Version 2.0
Document Number GC31-8496-06
Note |
---|
Before using this information and the product it supports, be sure to read the general information under Appendix I, Notices. |
Seventh Edition (September 2001)
© Copyright International Business Machines Corporation 2001. All rights reserved.
U.S. Government Users Restricted Rights -- Use, duplication or disclosure restricted by GSA ADP Schedule Contract with IBM Corp.
Introducing Network Dispatcher
Planning for the Dispatcher component
Configuring the Dispatcher component
Planning for the Content Based Routing component
Configuring the Content Based Routing component
Planning for the Mailbox Locator component
Configuring the Mailbox Locator component
Planning for the Site Selector component
Configuring the Site Selector component
Planning for the Consultant for Cisco CSS Switches component
Configuring the Consultant for Cisco CSS Switches component
Advanced Network Dispatcher Functions
Operating and managing Network Dispatcher
Appendix A. How to read a syntax diagram
Appendix B. Command reference for Dispatcher, CBR, and Mailbox Locator
Appendix C. Content rule (pattern) syntax
Appendix D. Command reference for Site Selector
Appendix E. Command reference for Consultant for Cisco CSS Switches
Appendix F. Sample configuration files
This book explains how to plan for, install, configure, use, and troubleshoot IBM(R) WebSphere(TM) Edge Server Network Dispatcher for AIX, Linux, Solaris, and Windows 2000. Previously, this product was called SecureWay Network Dispatcher, eNetwork Dispatcher, and Interactive Network Dispatcher.
The most current version of this book is available in HTML and PDF formats on the WebSphere Edge Server Web site. To access the online book, go to the following URL:
http://www.ibm.com/software/webservers/edgeserver/library.html
The WebSphere Edge Server Web site gives you the latest details on how you can use Network Dispatcher to maximize the performance of your servers. Configuration examples and scenarios are included. To access this Web site, go to the following URL:
http://www.ibm.com/software/webservers/edgeserver
For the most current updates and usage hints about Network Dispatcher, visit the WebSphere Edge Server support Web page and click Search for Network Dispatcher hints and tips. To access this Web page, go to the following URL:
http://www.ibm.com/software/webservers/edgeserver/support.html
Your feedback is important in helping to provide the most accurate and high-quality information. If you have any comments about this book or any other WebSphere Edge Server document:
How quickly can you make Network Dispatcher work for you? Consider the following:
Assume you're the webmaster for the Intersplash Corporation. You manage a local Web site with two HTTP servers. You've been using a round-robin approach to manage the load on the two servers, but business has picked up recently and customers are starting to complain about not being able to access the site. What do you do?
Go out to http://www.ibm.com/software/webservers/edgeserver and download the latest version of Network Dispatcher. This product has five components: Dispatcher, Content Based Routing (CBR), Mailbox Locator, Site Selector, and Consultant for Cisco CSS Switches (Cisco Consultant). For the time being, we'll just discuss the Dispatcher component.
Figure 1. A simple local Dispatcher configuration
This quick-start example shows how to configure three locally-attached workstations using the Dispatcher component's MAC forwarding method to load-balance Web traffic between two Web servers. The configuration would be essentially the same for balancing any other TCP or stateless UDP application traffic.
For the quick-start example, you will need three workstations and four IP addresses. One workstation will be used as the Dispatcher; the other two workstations will be used as Web servers. Each Web server requires one IP address. The Dispatcher workstation requires one actual address, and one address to be load balanced.
Workstation | Name | IP Address |
---|---|---|
1 | server1.intersplash.com | 9.67.67.101 |
2 | server2.intersplash.com | 9.67.67.102 |
3 | server3.intersplash.com | 9.67.67.103 |
Netmask = 255.255.255.0 |
Each of the workstations contains only one standard Ethernet network interface card.
Name= www.intersplash.com IP=9.67.67.104
Add an alias for www.intersplash.com to the loopback interface on server2.intersplash.com and server3.intersplash.com.
ifconfig lo0 alias www.intersplash.com netmask 255.255.255.0
ifconfig lo0:1 www.intersplash.com 127.0.0.1 up
You have now completed all configuration steps that are required on the two Web server workstations.
With Dispatcher, you can create a configuration by using the command line, the configuration wizard, or the graphical user interface (GUI).
If you are using the command line, follow these steps:
ndcontrol executor start
ndcontrol cluster add www.intersplash.com
ndcontrol port add www.intersplash.com:80
ndcontrol server add www.intersplash.com:80:server2.intersplash.com
ndcontrol server add www.intersplash.com:80:server3.intersplash.com
ndcontrol cluster configure www.intersplash.com
ndcontrol manager start
Dispatcher will now do load balancing based on server performance.
ndcontrol advisor start http 80
Dispatcher will now make sure that client requests are not sent to a failed Web server.
Your basic configuration with locally attached servers is now complete.
If you are using the configuration wizard, follow these steps:
ndserver
The wizard guides you step by step through the process of creating a basic configuration for the Dispatcher component. You will be asked questions about your network. You will be guided through the setup of a cluster for Dispatcher to load balance traffic between a group of servers.
With the configuration wizard, you will see the following panels:
Figure 2. The graphical user interface (GUI)
To start the graphical user interface, follow these steps:
ndserver
The left side of the panel displays a tree structure with Network Dispatcher at the top level, and Dispatcher, Content Based Routing, Mailbox Locator, Site Selector, and Cisco Consultant as components. See Figure 2.
All of the components can be configured from the GUI. You can select elements in the tree structure by clicking mouse button one (normally the left button) and then display pop-up menus by clicking mouse button two (normally the right button). The pop-up menus for the tree elements are also accessible from the menu bar located at the top of the panel.
Click the plus or minus signs to expand or compact the items in the tree structure.
The right side of the panel displays status indicator tabs for the element currently selected.
To access Help, click the question mark in the upper right hand corner of the Network Dispatcher window.
Test to see if the configuration is working.
There are many ways that you can configure Network Dispatcher to support your site. If you have only one host name for your site to which all of your customers will connect, you can define a single cluster of servers. For each of these servers, you configure a port through which Network Dispatcher communicates. See Figure 3.
Figure 3. Example of Dispatcher configured with a single cluster and 2 ports
In this example for the Dispatcher component, one cluster is defined at www.productworks.com. This cluster has two ports: port 80 for HTTP and port 443 for SSL. A client making a request to http://www.productworks.com (port 80) would go to a different server than a client requesting https://www.productworks.com (port 443).
Another way of configuring Network Dispatcher would be appropriate if you have a very large site with many servers dedicated to each protocol supported. In this case, you might want to define a cluster for each protocol with a single port but with many servers, as shown in Figure 4.
Figure 4. Example of Dispatcher configured with two clusters, each with one port
In this example for the Dispatcher component, two clusters are defined: www.productworks.com for port 80 (HTTP) and www.testworks.com for port 443 (SSL).
A third way of configuring Network Dispatcher would be necessary if your site does content hosting for several companies or departments, each one coming into your site with a different URL. In this case, you might want to define a cluster for each company or department and then define any ports to which you want to receive connections at that URL, as shown in Figure 5.
Figure 5. Example of Dispatcher configured with 2 clusters, each with 2 ports
In this example for the Dispatcher component, two clusters are defined with port 80 for HTTP and port 23 for Telnet for each of the sites at www.productworks.com and www.testworks.com.
This chapter instructs you on the hardware requirements and installation of Network Dispatcher on AIX, Linux, Solaris, and Windows 2000. Follow these instructions beginning at:
Notes:
To ensure that the Network Dispatcher components use the correct version of Java when multiple versions are installed, do the following:
Edit the script files for each component of Network Dispatcher that you are upgrading. The script files for each component are:
For example, on Windows 2000, if Java 1.3 is installed in C:\Program Files\IBM\Java13\jre\bin, change the line in ndserver.cmd:
IBM AIX 4.3.3.10 plus apars (in order to support Java 1.3). Refer to the README for the IBM AIX Developer Kit for a list of required AIX apars.
Table 1 lists the installp images for Network Dispatcher for
AIX.
Dispatcher (component, administration, license, and messages) | intnd.nd.driver intnd.nd.rte intnd.msg.nd.<language>.nd intnd.admin.rte intnd.msg.<language>.admin |
Administration (only) | intnd.admin.rte intnd.msg.<language>.admin |
Documentation | intnd.doc.rte |
License | intnd.nd.license |
Metric Server | intnd.ms.rte |
where <language> is one of:
If you are downloading an evaluation copy of the product from the Web site, use the installation instructions on (http://www.ibm.com/software/webservers/edgeserver/download.html).
When you install the product, you are given the option of installing any or all of the following:
Follow these steps to install Network Dispatcher for AIX:
Using SMIT:
When the command completes, press Done, and then select Exit Smit from the Exit menu or press F12. If using SMITTY, press F10 to exit the program.
Using the Command Line:
If installing from a CD, you must enter the following commands to mount the CD:
mkdir /cdrom mount -v cdrfs -p -r /dev/cd0 /cdrom
Refer to the following table to determine which command(s) to enter to
install the desired Network Dispatcher packages for AIX:
Network Dispatcher (with msgs). Includes: Dispatcher, CBR, Mailbox Locator, Site Selector, and Cisco Consultant | installp -acXgd device intnd.nd.rte intnd.admin.rte intnd.nd.driver intnd.msg.<language>.nd intnd.msg.<language>.admin |
Documents | installp -acXgd device intnd.doc.rte intnd.msg.<language>.doc |
Administration (only) | installp -acXgd device intnd.admin.rte intnd.msg.<language>.admin |
License | installp -acXgd device intnd.nd.license |
Metric Server | installp -acXgd device intnd.ms.rte intnd.msg.<language>.admin |
where device is:
Ensure that the result column in the summary contains SUCCESS for each part of Network Dispatcher that you are installing (APPLYing). Do not continue until all of the parts you wish to install are successfully applied.
installp -ld device
where device is:
To unmount the CD, type:
unmount /cdrom
lslpp -h | grep intnd
If you installed the full product, this command returns the following:
intnd.admin.rte intnd.doc.rte intnd.ms.rte intnd.msg.en_US.admin.rte intnd.msg.en_US.doc intnd.msg.en_US.nd.rte intnd.nd.driver intnd.nd.license intnd.nd.rte
Network Dispatcher install paths include the following:
This section explains how to install Network Dispatcher on Red Hat Linux or SuSE Linux using the product CD or the downloaded evaluation copy of the product from the Web site. Installation instructions can be found on the Web site (http://www.ibm.com/software/webservers/edgeserver/download.html).
Before beginning the installation procedure, ensure that you have root authority to install the software.
To install Network Dispatcher:
The installation image is a file in the format ndlinux-version.tar.
The following is a list of the RPM installable packages.
The command to install the packages should be issued from the same directory where the RPM files reside. Issue the following command to install each package: rpm -i package.rpm.
rpm -i --nodeps package.rpm
rpm -qa | grep ibmnd
Installing the full product should produce a listing like the following:
For Solaris 8, Netscape Navigator 4.07 (or higher) or Netscape Communicator 4.61 (or higher) for viewing online Help
This section explains how to install Network Dispatcher on Solaris using the product CD. If you are downloading an evaluation copy of the product from the internet, use the installation instructions on the Web site (http://www.ibm.com/software/webservers/edgeserver/download.html).
Before beginning the installation procedure, ensure that you have root authority to install the software.
To install Network Dispatcher:
At the command prompt, enter pkgadd -d pathname, where -d pathname is the device name of the CD-ROM drive or the directory on the hard drive where the package is located; for example: pkgadd -d /cdrom/cdrom0/.
You will be given a list of packages to install. They are:
If you want to install all of the packages, simply type "all" and press return. If you want to install some of the components, enter the name(s) corresponding to the packages to be installed separated by a space or comma and press return. You may be prompted to change permissions on existing directories or files. Simply press return, or answer "yes." You need to install prerequisite packages (because it installs in alphabetical order not prerequisite order). If you say "all" then just answer "yes" to all prompting and the install will complete successfully.
All of the packages depend on the common package, ibmndadm. This common package must be installed along with any of the other packages.
If you want to install the entire Network Dispatcher product, you must install five pieces: ibmdsp, ibmdsplic, ibmndadm, ibmnddoc, and ibmndms. If you want to install the remote administration , you only have to install one piece: ibmndadm.
The Network Dispatcher components reside in /opt/nd/servers install directory.
Installing the full product should produce a listing like the following:
You must download both the Developer Kit installable package and the Runtime Environment installable package prior to running the InstallShield program. (For information on running multiple versions of Java, see Note number (JVER).)
This section explains how to install Network Dispatcher on Windows 2000 using the product CD. If you are downloading an evaluation copy of the product from the Web site, use the installation instructions on the Web site (http://www.ibm.com/software/webservers/edgeserver/download.html).
You will be given a choice of packages to install.
They are:
Windows 2000 version of Network Dispatcher is supported on the following:
Restrictions: The Windows 2000 version of Network Dispatcher cannot be installed on the same machine with IBM Firewall.
Before beginning the installation procedure, ensure you have logged in as the Administrator or as a user with administrative privileges.
If you have an earlier version installed, you should uninstall that copy before installing the current version. To uninstall using the Add/Remove Program, do the following:
To install Network Dispatcher:
E:\setup
Network Dispatcher install paths include the following:
This chapter gives an overview of Network Dispatcher and includes the following sections:
Network Dispatcher is a software solution for load-balancing servers. It boosts the performance of servers by directing TCP/IP session requests to different servers within a group of servers; in this way, it balances the requests among all the servers. This load balancing is transparent to users and other applications. Network Dispatcher is useful for applications such as e-mail servers, World Wide Web servers, distributed parallel database queries, and other TCP/IP applications.
When used with Web servers, Network Dispatcher can help maximize the potential of your site by providing a powerful, flexible, and scalable solution to peak-demand problems. If visitors to your site can't get through at times of greatest demand, use Network Dispatcher to automatically find the optimal server to handle incoming requests, thus enhancing your customers' satisfaction and your profitability.
Network Dispatcher consists of five components that can be used separately or together to provide superior load-balancing results:
For HTTP protocol, you can also use the Dispatcher's content-based routing feature to load balance based on the content of the client request. The chosen server is the result of matching the URL to a specified rule.
For more information on the Dispatcher, CBR, Mailbox Locator, Site Selector, and Consultant for Cisco CSS Switches components, see What are the components of Network Dispatcher?.
The number of users and networks connected to the global Internet is growing exponentially. This growth is causing problems of scale that can limit users' access to popular sites.
Currently, network administrators are using numerous methods to try to maximize access. Some of these methods allow users to choose a different server at random if an earlier choice is slow or not responding. This approach is cumbersome, annoying, and inefficient. Another method is standard round-robin, in which the domain name server selects servers in turn to handle requests. This approach is better, but still inefficient because it blindly forwards traffic without any consideration of the server workload. In addition, even if a server fails, requests continue to be sent to it.
The need for a more powerful solution has resulted in Network Dispatcher. It offers numerous benefits over earlier and competing solutions:
As the number of client requests increases, you can add servers dynamically, providing support for tens of millions of requests per day, on tens or even hundreds of servers.
Load balancing ensures that each group of servers makes optimum use of its hardware by minimizing the hot-spots that frequently occur with a standard round-robin method.
Network Dispatcher uses standard TCP/IP protocols. You can add it to your existing network without making any physical changes to the network. It is simple to install and configure.
Using simple MAC level forwarding method, Dispatcher needs only to look at the inbound client-to-server flows. It does not need to see the outbound server-to-client flows. This significantly reduces its impact on the application compared with other approaches and can result in improved network performance.
Dispatcher offers built-in high availability, utilizing a backup machine that remains ready at all times to take over load balancing should the primary Dispatcher machine fail. Dispatcher also offers mutual high availability which allows two machines to be both active and standby for each other. See How about high availability?.
In conjunction with Caching Proxy, the CBR component has the ability to proxy HTTP and HTTPS (SSL) requests to specific servers based on the content requested. For example, if a request contains the string "/cgi-bin/" in the directory portion of the URL, and the server name is a local server, CBR can direct the request to the best server in a set of servers specifically allocated to handle cgi requests.
The Dispatcher component also provides content-based routing, but it does not require the Caching Proxy to be installed. Because the Dispatcher component's content-based routing is performed in the kernel as packets are received, it can provide faster content-based routing than the CBR component. The Dispatcher component performs content-based routing for HTTP (using the "content" type rule) and HTTPS (using SSL session ID affinity).
Network Dispatcher for IBM WebSphere Edge Server Version 2.0 contains a number of new features. The most significant are listed here.
This feature applies to all the Network Dispatcher components.
Network Dispatcher now supports a newer version of AIX: AIX v5.1. See Requirements for AIX for more information.
This feature applies to all the Network Dispatcher components.
Network Dispatcher now supports SuSE Linux v7.1 (kernel version 2.4.0-4GB). Previously, Network Dispatcher only supported Red Hat Linux. See Requirements for Red Hat Linux or SuSE Linux for more information.
This feature applies to all the Network Dispatcher components.
Network Dispatcher now supports a newer version of Red Hat Linux: Red Hat Linux v7.1 (kernel version 2.4.2-2). See Requirements for Red Hat Linux or SuSE Linux for more information.
This feature applies to all the Network Dispatcher components.
On Linux and Solaris operating systems, Network Dispatcher offers NLS for the Group 1 countries.
This feature applies to all the Network Dispatcher components.
Network Dispatcher provides NLS for the new Chinese Standard GB 18030.
This feature is a new component for Network Dispatcher.
Collaboration with Cisco and their Content Distribution Network (CDN) has led to the development of an additional component for Network Dispatcher -- Cisco Consultant. This component (which was first introduced as a standalone Preview) allows the Network Dispatcher to generate weights and make load balancing decisions for the Cisco CSS Switch.
See Planning for the Consultant for Cisco CSS Switches component and Configuring the Consultant for Cisco CSS Switches component for more information.
This feature is a new component for Network Dispatcher.
The Site Selector component balances the load among a group of servers by selecting the "right" server's IP address for a name service request. This allows the client to connect directly to the server for all its communication. Site Selector replaces Interactive Session Support (ISS), a Network Dispatcher component in prior releases. Site Selector provides similar functionality as ISS but requires fewer steps when setting up a typical DNS load balancing configuration.
See Planning for the Site Selector component and Configuring the Site Selector component for more information.
This feature applies to all the Network Dispatcher components.
Metric Server provides server load information to the Network Dispatcher in the form of system-specific metrics. Metric Server agent is a component of Network Dispatcher that can be installed and run on servers that Network Dispatcher is load balancing. Metric Server replaces the System Monitoring Agent (SMA), which was supported on Linux, in previous releases. Metric Server is supported on all platforms. It is recommended that Metric Server be used in conjunction with the Site Selector component.
See Metric Server for more information.
This feature is a new component for Network Dispatcher.
The Mailbox Locator component was formerly a feature within the CBR component that load balanced across IMAP and POP3 mail servers based on userID and password. Separating CBR into two components permits Mailbox Locator (formerly known as "CBR for IMAP/POP3") and CBR with Caching Proxy to be run on the same machine.
See Planning for the Mailbox Locator component and Configuring the Mailbox Locator component for more information.
Configuring Caching Proxy configuration file (ibmproxy.conf) to use CBR has been streamlined and CBR has been enhanced so that multiple instances of the Caching Proxy can run simultaneously on the same machine while interfacing with CBR. For more information on how to configure CBR with Caching Proxy, see Setting up the CBR machine.
This feature applies to the Dispatcher component.
NAT/NAPT removes the limitation for backend servers to be located on a locally attached network. It also enables Dispatcher to load balance the client's TCP requests to multiple server daemons running on the same physical machine. You can configure servers with multiple daemons in two different ways. With NAT, you can configure multiple server daemons to respond to requests to different IP addresses. This is also known as binding a server daemon to an IP address. With NAPT, you can configure multiple server daemons to listen on different port numbers.
The advantage of Dispatcher's nat forwarding method is that it is configured at the port level giving you much better granularity.
See Dispatcher's NAT/NAPT (nat forwarding method) for more information.
This feature applies to the Dispatcher component.
In prior Network Dispatcher releases, content-based routing was only available using the CBR component in conjunction with Caching Proxy. The Dispatcher component now allows you to perform content-based routing for HTTP (using the "content" type rule) and HTTPS (using SSL session ID affinity) without Caching Proxy. For HTTP and HTTPS traffic, the Dispatcher component can provide faster content-based routing than the CBR component.
See Dispatcher's content-based routing (cbr forwarding method) for more information on using the content rule and SSL session ID affinity.
This feature applies to the Dispatcher component's content-based routing feature (cbr forwarding method) and the CBR component.
Passive cookie affinity allows you to load-balance Web traffic with affinity to the same server based upon self-identifying cookies generated by the servers. See Passive cookie affinity for more information.
This feature applies to the Dispatcher component's content-based routing feature (cbr forwarding method) and the CBR component.
URI affinity allows you to load-balance Web traffic to caching-proxy servers in a manner that effectively increases the size of the cache. See URI affinity for more information.
This feature applies to all Network Dispatcher components.
In earlier releases, the proportion of importance (given to active connections, new connections, port, and system metrics) for determining load-balancing decisions was set from the manager function. These proportions were applied to every cluster in the configuration for the component. All clusters were measured using the same proportions regardless of the site it was load balancing.
With this enhancement, you can set the proportion of importance on a per cluster (or site) basis. See Proportion of importance given to status information for more information.
This feature applies to all the Network Dispatcher components.
Network Dispatcher now provides the ability to partition one physical server into several logical servers. This allows you to query, for example, a particular service on the machine to detect if a servlet engine or a database request is running faster, or not running at all. This enhancement provides you with the capability to distribute load based on more granular service-specific workload. See Server Partitioning: logical servers configured to one physical server (IP address) for more information.
This feature applies to the Dispatcher and CBR components.
With this enhancement for the HTTP advisor, you can assess the health of individual services within a server. For each logical server under the HTTP port, you can specify a unique client HTTP URL string, specific for the service that you want to query on the server. See HTTP advisor request/response (URL) option for more information.
This feature applies to all the Network Dispatcher components.
Network Dispatcher allows you to start different advisors running on the same port but configured on different clusters (sites). For example, this feature will allow you to use an HTTP advisor on port 80 for one cluster (site) and a custom advisor on port 80 for another cluster (site). See Starting and stopping an advisor for more information.
This feature applies to the Dispatcher component.
With this enhancement, Dispatcher provides the ability to detect potential denial of service attacks and notify administrators via an alert. Dispatcher does this by analyzing incoming requests for a conspicuous amount of half-open connections, a common trait of simple denial of service attacks.
See Denial of service attack detection for more information.
This feature applies to all components except Consultant for Cisco CSS Switches and Site Selector.
Network Dispatcher provides additional user exits that trigger scripts which you can customize. You can create the scripts to perform automated actions, such as logging when a high availability state has changed or alerting an Administrator when servers are marked down. Network Dispatcher provides the following new sample script files:
This feature applies to the Dispatcher component.
Dispatcher provides a DB2 advisor which communicates with the DB2 servers. See List of advisors for more information on the DB2 advisor.
The five components of Network Dispatcher are: Dispatcher, Content Based Routing (CBR), Mailbox Locator, Site Selector, and Consultant for Cisco CSS Switches. Network Dispatcher gives you the flexibility of using the components separately or together depending on your site configuration. This section gives an overview of the these components.
The Dispatcher component balances traffic among your servers through a unique combination of load balancing and management software. Dispatcher can also detect a failed server and forward traffic around it. Dispatcher supports HTTP, FTP, SSL, SMTP, NNTP, IMAP, POP3, Telnet, and any other TCP or stateless UDP based application.
All client requests sent to the Dispatcher machine are directed to the "best" server according to weights that are set dynamically. You can use the default values for those weights or change the values during the configuration process.
Dispatcher offers three forwarding methods (specified on the port):
The Dispatcher component is the key to stable, efficient management of a large, scalable network of servers. With Dispatcher, you can link many individual servers into what appears to be a single, virtual server. Your site thus appears as a single IP address to the world. Dispatcher functions independently of a domain name server; all requests are sent to the IP address of the Dispatcher machine.
Dispatcher brings distinct advantages in balancing traffic load to clustered servers, resulting in stable and efficient management of your site.
Figure 6. Example of a physical representation of a site using Dispatcher to manage local servers
Figure 6 shows a physical representation of the site using an Ethernet network configuration. The Dispatcher machine can be installed without making any physical changes to the network. After a client request is directed to the optimal server by the Dispatcher, the response is then sent directly from server to client with no involvement by the Dispatcher when using MAC forwarding method.
Figure 7. Example of a site using Dispatcher and Metric Server to manage servers
Figure 7 illustrates a site in which all servers are on a local network. The Dispatcher component is used to forward requests and the Metric Server is used to provide system load information to the Dispatcher machine.
In this example, the Metric Server daemon is installed on each backend server. You can use Metric Server with the Dispatcher component or any of the other Network Dispatcher components.
Figure 8. Example of a site using Dispatcher to manage local and remote servers
Wide area support in Dispatcher enables you to use both local and remote servers (servers on different subnets). Figure 8 shows a configuration where one local Dispatcher (Dispatcher 1) serves as the entry point for all requests. It distributes these requests among its own local servers (ServerA, ServerB, ServerC) and to the remote Dispatcher (Dispatcher 2), which will load balance to its local servers (ServerG, ServerH, ServerI).
When using Dispatcher's NAT forwarding method or using GRE support, wide area support with Dispatcher can also be achieved without using a Dispatcher at the remote site (where ServerD, ServerE, and ServerF are located). See Dispatcher's NAT/NAPT (nat forwarding method) and GRE (Generic Routing Encapsulation) support for more information.
CBR works with Caching Proxy to proxy client requests to specified HTTP or HTTPS (SSL) servers. It allows you to manipulate caching details for faster Web document retrieval with low network bandwidth requirements. CBR along with Caching Proxy examines HTTP requests using specified rule types.
CBR gives you the ability to specify a set of servers that should handle a request based on regular expression matching of the content of the request. Because CBR allows you to specify multiple servers for each type of request, the requests can be load balanced for optimal client response. CBR will also detect when one server in a set has failed, and stop routing requests to that server. The load balancing algorithm used by the CBR component is identical to the proven algorithm used by the Dispatcher component.
When a request is received by Caching Proxy, it is checked against the rules that have been defined in the CBR component. If a match is found, then one of the servers associated with that rule is chosen to handle the request. Caching Proxy then performs its normal processing to proxy the request to the chosen server.
CBR has the same functions as Dispatcher with the exception of high availability, subagent, wide area, and a few other configuration commands.
Caching Proxy must be running before CBR can begin load balancing client requests.
Figure 9. Example of a site using CBR to manage local servers
Figure 9 shows a logical representation of a site in which CBR is being used to proxy some content from local servers. The CBR component uses Caching Proxy to forward client requests (HTTP or HTTPS) to the servers based on the content of the URL.
Mailbox Locator can provide a single point of presence for many IMAP or POP3 servers. Each server can have a subset of all user mailboxes serviced by the point of presence. For IMAP and POP3 traffic, Mailbox Locator is a proxy that chooses an appropriate server based on userID and password provided by the client. Mailbox Locator does not support rules-based load balancing.
Figure 10. Example of a site using Mailbox Locator to manage local servers
Figure 10 shows a logical representation of a site in which Mailbox Locator is being used to proxy client requests (IMAP or POP3 protocol) to the appropriate server, based on userID and password.
Site Selector acts as a name server that works in conjunction with other name servers in a domain name system to load balance among a group of servers using measurements and weights that are gathered. You can create a site configuration to let you load balance traffic among a group of servers based on the domain name used for a client's request.
A client submits a request for resolution of a domain name to a name server within its network. Name server forwards the request to the Site Selector machine. Site Selector then resolves the domain name to the IP address of one of the servers that has been configured under the site name. Site Selector returns the IP address of the selected server to the name server. Name server returns the IP address to the client.
Metric Server is a system monitoring component of Network Dispatcher that must be installed in each load-balanced server within your configuration. Using Metric Server, Site Selector can monitor the level of activity on a server, detect when a server is the least heavily loaded, and detect a failed server. The load is a measure of how hard the server is working. By customizing system metric script files, you can control the type of measurements used to measure the load. You can configure Site Selector to suit your environment, taking into account such factors as frequency of access, the total number of users, and types of access (for example, short queries, long-running queries, or CPU-intensive loads).
Figure 11 illustrates a site in which the Site Selector component is used to answer requests. Server1, Server2, and Server3 are local. Server4, Server5, and Server6 are remote.
A client submits a request for resolution of a domain name to a client name server. The client name server forwards the request through the DNS to the Site Selector machine (Path 1). Site Selector then resolves the domain name to the IP address of one of the servers. Site Selector returns the IP address of the selected server to the client name server. The name server returns the IP address to the client.
Once the client receives the IP address of the server, the client routes subsequent requests directly to the selected server (Path 2).
Consultant for Cisco CSS Switches forms a complementary solution in conjunction with Cisco's CSS 11000 series switches. The combined solution blends the CSS 11000 series' robust packet forwarding and content routing capabilities with Network Dispatcher's sophisticated awareness algorithms for determining backend server, application, and database availability and loading information. The Cisco Consultant function utilizes Network Dispatcher's manager, standard and custom advisors, and Metric Server to determine the metrics, health, and loading of the backend servers, applications, and databases. With this information Cisco Consultant generates server weighting metrics, which it sends to the Cisco CSS Switch for optimal server selection, load optimization, and fault tolerance.
The Cisco CSS Switch makes load-balancing decisions based on user-specified criteria.
Cisco Consultant tracks many criteria, including:
When a Cisco CSS Switch, without Cisco Consultant, is determining the health of a content-providing server, it uses response times for content requests or other network measures. With Cisco Consultant in place, these activities are offloaded from the Cisco CSS Switch to Cisco Consultant. Cisco Consultant influences the server's weight or ability to serve content, and activates or suspends a server as appropriate when the server regains or loses availability.
Cisco Consultant:
Weights are applied to all servers on a port. For any particular port, the requests are distributed between servers based on their weights relative to each other. For example, if one server is set to a weight of 10, and the other to 5, the server set to 10 should get twice as many requests as the server set to 5. These weights are provided to the Cisco CSS Switch using SNMP. As the weight of any server is set higher, the Cisco CSS Switch directs more requests to that server.
Figure 12. Example of a site using Cisco Consultant and Metric Server to manage local servers
Cisco Consultant, in conjunction with the Cisco CSS Switch, delivers a "best of both worlds" solution that combines wire-speed content switching with sophisticated application awareness, fault tolerance, and server load optimization. Cisco Consultant is part of an overall complementary solution between the Cisco CSS Switch and IBM's WebSphere Edge Server.
Refer to Installing Network Dispatcher for a list of Cisco Consultant requirements.
The Dispatcher component offers a built-in high availability feature. This feature involves the use of a second Dispatcher machine that monitors the main, or primary, machine and stands by to take over the task of load balancing should the primary machine fail at any time. The Dispatcher component also offers mutual high availability which allows two machines to be both primary and secondary (backup) for each other. See Configure high availability.
When using a two-tier configuration with a Dispatcher machine load balancing traffic to two or more server machines that have either CBR, Mailbox Locator or Site Selector, you can achieve a level of high availability for these components of Network Dispatcher.
This chapter describes what the network planner should consider before installing and configuring the Dispatcher component.
This chapter includes the following sections:
Platform requirements:
Dispatcher consists of the following functions:
Using the manager is optional. However, if the manager is not used, load balancing will be performed using weighted round-robin scheduling based on the current server weights, and advisors will not be available.
Dispatcher also offers advisors that do not exchange protocol-specific information, such as: the DB2 advisor that reports on the health of DB2 servers and the Ping advisor that reports whether the server responds to a ping. For a complete list of advisors, see List of advisors.
You also have the option of writing your own advisors (see Create custom (customizable) advisors).
Using the advisors is optional but recommended.
The three key functions of Dispatcher (executor, manager, and advisors) interact to balance and dispatch the incoming requests between servers. Along with load balancing requests, the executor monitors the number of new connections, active connections, and connections in a finished state. The executor also does garbage collection of completed or reset connections and supplies this information to the manager.
The manager collects information from the executor, the advisors, and a system-monitoring program, such as Metric Server. Based on the information the manager receives, it adjusts how the server machines are weighted on each port and gives the executor the new weighting for use in its balancing of new connections.
The advisors monitor each server on the assigned port to determine the server's response time and availability and then give this information to the manager. The advisors also monitor whether a server is up or down. Without the manager and the advisors, the executor does round-robin scheduling based on the current server weights.
Figure 13. Example of a Dispatcher using simple high availability
The high availability feature involves the use of a second Dispatcher machine. The first Dispatcher machine performs load balancing for all the client traffic as it would in a single Dispatcher configuration. The second Dispatcher machine monitors the "health" of the first, and will take over the task of load balancing if it detects that the first Dispatcher machine has failed.
Each of the two machines is assigned a specific role, either primary or backup. The primary machine sends connection data to the backup machine on an ongoing basis. While the primary is active (load balancing), the backup is in a standby state, continually updated and ready to take over, if necessary.
The communication sessions between the two machines are referred to as heartbeats. The heartbeats allow each machine to monitor the health of the other.
If the backup machine detects that the active machine has failed, it will take over and begin load balancing. At that point the statuses of the two machines are reversed: the backup machine becomes active and the primary becomes standby.
In the high availability configuration, both primary and backup machines must be on the same subnet.
For information about configuring high availability, see High availability.
Figure 14. Example of a Dispatcher using mutual high availability
The mutual high availability feature involves the use of two Dispatcher machines. Both machines actively perform load balancing of client traffic, and both machines provide backup for each other. In a simple high availability configuration, only one machine performs load balancing. In a mutual high availability configuration, both machines load balance a portion of the client traffic.
For mutual high availability, client traffic is assigned to each Dispatcher machine on a cluster address basis. Each cluster can be configured with the NFA (nonforwarding address) of its primary Dispatcher. The primary Dispatcher machine normally performs load balancing for that cluster. In the event of a failure, the other machine performs load balancing for both its own cluster and for the failed Dispatcher's cluster.
For an illustration of a mutual high availability configuration with shared "cluster set A" and shared "cluster set B," see Figure 14. Each Dispatcher can actively route packets for its primary cluster. If either Dispatcher were to fail and could no longer actively route packets for its primary cluster, then the other Dispatcher could take over routing packets for its backup cluster.
For information about configuring high availability and mutual high availability, see High availability.
Using Dispatcher's MAC forwarding method (the default forwarding method), Dispatcher load balances the incoming request to the selected server and the server returns the response directly to the client without any involvement of the Dispatcher. With this forwarding method, Dispatcher only looks at the inbound client-to-server flows. It does not need to see the outbound server-to-client flows. This significantly reduces its impact on the application and can result in improved network performance.
The forwarding method can be selected when adding a port using the ndcontrol port add cluster:port method value command. The default forwarding method value is mac. You can specify the method parameter only when the port is added. Once you add the port, you cannot change the setting of the forwarding method. See ndcontrol port -- configure ports for more information.
Using Dispatcher's Network Address Translation (NAT) or Network Address Port Translation (NAPT) capability removes the limitation for load-balanced servers to be located on a locally attached network. When you want to have servers located at remote locations, you can use the NAT forwarding method technique rather than using a GRE/WAN encapsulation technique. You can also use the NAPT feature to access multiple server daemons residing on each load-balanced server machine, where each daemon listens on a unique port.
You can configure a server with multiple daemons in two different ways:
This application works well with upper-level application protocols such as HTTP, SSL, IMAP, POP3, NNTP, SMTP, Telnet, etc.
Limitations:
To implement NAT/NAPT:
ndcontrol server add cluster:port:server mapport value returnaddress rtrnaddress router rtraddress
This will map the client request's destination port number (which is for Dispatcher) to the server's port number that Dispatcher uses to load balance the client's request. Mapport allows Network Dispatcher to receive a client's request on one port and to transmit it to a different port on the server machine. With mapport you can load balance a client's requests to a server machine that may have multiple server daemons running. The default for mapport is the client request's destination port number.
The return address is a unique address or hostname that you configure on the Dispatcher machine. Dispatcher uses the return address as its source address when load balancing the client's request to the server. This ensures that the server will return the packet to the Dispatcher machine rather than sending the packet directly to the client. (Dispatcher will then forward the IP packet to the client.) You must specify the return address value when adding the server. You cannot modify the return address unless you remove the server and then add it again. The return address cannot be the same as the cluster, server, or NFA address.
The address of the router to the remote server.
In prior Network Dispatcher releases, content-based routing was only available using the CBR component in conjunction with Caching Proxy. The Dispatcher component now allows you to perform content-based routing for HTTP (using the "content" type rule) and HTTPS (using SSL session ID affinity) without Caching Proxy. For HTTP and HTTPS traffic, the Dispatcher component can provide faster content-based routing than the CBR component.
For HTTP: Server selection, for Dispatcher's content-based routing, is based upon the contents of a URL or an HTTP header. It is configured using the "content" type rule. When configuring the content rule, specify the search string "pattern" and a set of servers to the rule. When processing a new incoming request, this rule compares the specified string with the client's URL or with the specified HTTP header in the client request.
If Dispatcher finds the string in the client request, Dispatcher forwards the request to one of the servers within the rule. Dispatcher then relays the response data from the server to the client ("cbr" forwarding method).
If Dispatcher does not find the string in the client request, Dispatcher does not select a server from the set of servers within the rule.
For HTTPS (SSL): Dispatcher's content-based routing load balances based on the SSL ID session field of the client request. With SSL, a client request contains the SSL session ID of a prior session, and servers maintain a cache of their prior SSL connections. Dispatcher's SSL ID session affinity allows the client and server to establish a new connection using the security parameters of the previous connection with the server. By eliminating the renegotiation of SSL security parameters, such as shared keys and encryption algorithms, the servers will save CPU cycles and the client will get a quicker response. In order to enable SSL session ID affinity, port stickytime must be set to a nonzero value. When stickytime has been exceeded, the client may be sent to a different server from the previous.
To implement Dispatcher's content-based routing (cbr forwarding method):
ndcontrol server add cluster:port:server mapport value returnaddress rtrnaddress router rtraddress
ndcontrol rule 125.22.22.03:80:contentRule1 type content pattern pattern
Where pattern specifies the pattern to be used for the content type rule. For more information on the content rule type, see Using rules based on the request content. For more information on valid expressions for pattern, see Appendix C, Content rule (pattern) syntax.
For HTTPS (SSL): To configure the SSL ID session affinity, set the stickytime parameter on the port to a nonzero value. For more information about stickytime on the port command, see ndcontrol rule -- configure rules.
Before following the steps in this chapter, see Planning for the Dispatcher component. This chapter explains how to create a basic configuration for the Dispatcher component of Network Dispatcher.
Table 3. Configuration tasks for the Dispatcher function
Task | Description | Related information |
---|---|---|
Set up the Dispatcher machine. |
Set up your load balancing configuration.
| Setting up the Dispatcher machine |
Set up machines to be load-balanced. | Alias the loopback device, check for an extra route, and delete any extra routes. | Setting up server machines for load balancing |
There are four basic methods of configuring the Dispatcher:
This is the most direct means of configuring the Dispatcher. The command parameter values must be entered in English characters. The only exceptions are host names (used in cluster, server, and highavailability commands) and file names (used in file commands).
To start Dispatcher from the command line:
You can enter a minimized version of the ndcontrol command parameters. You only need to enter the unique letters of the parameters. For example, to get help on the file save command, you can type ndcontrol he f instead of ndcontrol help file.
To start up the command line interface: issue ndcontrol to receive an ndcontrol command prompt.
To end the command line interface: issue exit or quit.
The commands for configuring the Dispatcher can be entered into a configuration script file and executed together. See Sample Network Dispatcher configuration files.
ndcontrol file appendload myscript
ndcontrol file newload myscript
For an example of the graphical user interface (GUI), see Figure 2.
To start the GUI, follow these steps
ndserver
In order to configure the Dispatcher component from the GUI, you must first select Dispatcher in the tree structure. You can start the executor and manager once you connect to a Host. You can also create clusters containing ports and servers, and start advisors for the manager.
The GUI can be used to do anything that you would do with the ndcontrol command. For example, to define a cluster using the command line, you would enter ndcontrol cluster add cluster command. To define a cluster from the GUI, right-click Executor, then in the pop-up menu, left-click Add Cluster. Enter the cluster address in the pop-up window, then click OK.
Pre-existing Dispatcher configuration files can be loaded using the Load New Configuration(for completely replacing the current configuration) and Append to Current Configuration (for updating the current configuration) options presented in the Host pop-up menu. You should save your Dispatcher configuration to a file periodically using the Save Configuration File As option also presented in the Host pop-up menu. The File menu located at the top of the GUI will allow you to save your current host connections to a file or restore connections in existing files across all Network Dispatcher components.
The configuration commands can also be run remotely. For more information, see Remote Authenticated Administration.
You can access Help by clicking the question mark icon in the upper right hand corner of the Network Dispatcher window.
For more information about using the GUI, see General Instructions for using the GUI.
For more information about using the configuration wizard, see Configuring using the configuration wizard.
Before setting up the Dispatcher machine, you must be the root user (for AIX, Linux, or Solaris) or the Administrator on Windows 2000.
For AIX, Linux, and Solaris only, the Network Dispatcher can have a collocated server. This simply means that the Network Dispatcher can physically reside on a server machine which it is load balancing.
You will need at least two valid IP addresses for the Dispatcher machine:
This IP address is the primary IP address of the Dispatcher machine and is called the nonforwarding address (NFA). This is by default the same address as that returned by the hostname command. Use this address to connect to the machine for administrative purposes, such as doing remote configuration via Telnet or accessing the SNMP subagent. If the Dispatcher machine can already ping other machines on the network, you do not need to do anything further to set up the nonforwarding address.
A cluster address is an address that is associated with a host name (such as www.yourcompany.com). This IP address is used by a client to connect to the servers in a cluster. This is the address that is load balanced by the Dispatcher.
Solaris Only:
For example, if you plan to use two 100Mbps Ethernet adapters, the ibmnd.conf file should have a single line specifying the hme device. If you plan to use one 10Mbps Ethernet adapter and one 100Mbps Ethernet adapter, you will have two lines in the ibmnd.conf file: one line specifying the le device and one line specifying the hme device.
The ibmnd.conf file provides input to the Solaris autopush command and must be compatible with the autopush command.
For example, if clusters X and Y are configured for use by the Mailbox Locator component on any of the adapters listed in ibmnd.conf, clusters X and Y are unconfigured when the ndcontrol executor start or ndcontrol executor stop commands are issued. This may not be the desired result. When clusters X and Y are configured in the goAliases script, the clusters are automatically reconfigured after the Dispatcher executor starts or stops.
Windows 2000 Only: Ensure that IP forwarding is not enabled for the TCP/IP protocol. (See your Windows 2000 TCP/IP configuration.)
Figure 15 shows an example of Dispatcher set up with a single cluster, two ports, and three servers.
Figure 15. Example of the IP addresses needed for the Dispatcher machine
For help with commands used in this procedure, see Appendix B, Command reference for Dispatcher, CBR, and Mailbox Locator.
For a sample configuration file, see Sample Network Dispatcher configuration files.
AIX, Linux, and Solaris: To start the server function, type ndserver.
Windows 2000: The server function starts automatically as a service.
To start the executor function, enter the ndcontrol executor start command. You may also change various executor settings at this time. See Appendix B, Command reference for Dispatcher, CBR, and Mailbox Locator.
The nonforwarding address is used to connect to the machine for administrative purposes, such as using Telnet or SMTP to this machine. By default, this address is the hostname.
To define the nonforwarding address, enter the ndcontrol executor set nfa IP_address command or edit the sample configuration file. IP_address is either the symbolic name or the dotted-decimal address.
Dispatcher will balance the requests sent to the cluster address to the servers configured on the ports for that cluster.
The cluster is either the symbolic name, the dotted decimal address, or the special address 0.0.0.0 that defines a wildcard cluster. To define a cluster, issue the command ndcontrol cluster add. To set cluster options, issue the command ndcontrol cluster set or you can use the GUI to issue commands. Wildcard clusters can be used to match multiple IP addresses for incoming packets to be load balanced. See Use wildcard cluster to combine server configurations, Use wildcard cluster to load balance firewalls, and Use wildcard cluster with Caching Proxy for transparent proxy for more information.
Once the cluster has been defined, you normally must configure the cluster address on one of the network interface cards of the Dispatcher machine. To do this, issue the command ndcontrol cluster configure cluster_address. This will look for an adapter with an existing address that belongs on the same subnet as the cluster address. It will then issue the operating system's adapter configuration command for the cluster address, using the adapter found and the netmask for the existing address found on that adapter. For example:
ndcontrol cluster configure 204.67.172.72
Circumstances where you would not want to configure the cluster address are for clusters added to a standby server in high-availability mode, or clusters added to a wide-area dispatcher acting as a remote server. You also do not need to run the cluster configure command if, in stand-alone mode, you use the sample goIdle script. For information on the goIdle script, see Using scripts.
In rare cases you may have a cluster address that does not match any subnet for existing addresses. In this case, use the second form of the cluster configure command and explicitly provide the interface name and netmask. Use ndcontrol cluster configure cluster_address interface_name netmask.
Some examples are:
ndcontrol cluster configure 204.67.172.72 en0 255.255.0.0 (AIX) ndcontrol cluster configure 204.67.172.72 eth0:1 255.255.0.0 (Linux) ndcontrol cluster configure 204.67.172.72 le0:1 255.255.0.0 (Solaris 7) ndcontrol cluster configure 204.67.172.72 le0 255.255.0.0 (Solaris 8) ndcontrol cluster configure 204.67.172.72 en0 255.255.0.0 (Windows 2000)
In order to use the second form of the cluster configure command on Windows 2000, you must determine the interface name to use.
If you have only one Ethernet card in your machine, the interface name will be en0. Likewise, if you have only one Token Ring card, the interface name will be tr0. If you have multiple cards of either type, you will need to determine the mapping of the cards. Use the following steps:
The network interface adapters are listed under Network Cards. Click each one to determine whether it is an Ethernet or Token Ring interface. The type of interface is listed in the Description column. The names assigned by ndconfig map to the interface types. For example, the first Ethernet interface in the list is assigned by ndconfig to en0, the second to en1, and so on; the first Token Ring interface is assigned to tr0, the second to tr1, and so on.
After you obtain this mapping information, you can create an alias on the network interface to the cluster address.
The cluster configure command merely runs ifconfig (or ndconfig on Windows 2000) commands, so you can still use the ifconfig (ndconfig) commands if you wish.
The ndconfig command is supplied with the Dispatcher component to configure cluster aliases using the command line. The ndconfig command has the same syntax as a UNIX ifconfig command.
ndconfig en0 alias 204.67.172.72 netmask 255.255.0.0
To determine the interface name, use the same technique as for the second form of the cluster configure command.
When using bind-specific server applications that bind to a list of IP addresses that do not contain the server's IP, use arp publish command instead of ifconfig to dynamically set an IP address on the Network Dispatcher machine. For example:
arp -s <cluster> <Network Dispatcher MAC address> pub
To define a port, enter the ndcontrol port add cluster:port command, edit the sample configuration file, or use the GUI. Cluster is either the symbolic name or the dotted-decimal address. Port is the number of the port you are using for that protocol. You may also change various port settings at this time. You must define and configure all servers for a port. See Appendix B, Command reference for Dispatcher, CBR, and Mailbox Locator.
Port number 0 (zero) is used to specify a wildcard port. This port will accept traffic for a port that is not destined for any of the defined ports on the cluster. The wildcard port will be used to configure rules and servers for any port. This function could also be used if you have an identical server/rule configuration for multiple ports. The traffic on one port could then affect the load-balancing decisions for traffic on other ports. See Use wildcard port to direct unconfigured port traffic for more information about when you might want to use a wildcard port.
To define a load-balanced server machine, enter the ndcontrol server add cluster:port:server command, edit the sample configuration file, or use the GUI. Cluster and server are either the symbolic name or the dotted-decimal address. Port is the number of the port you are using for that protocol. You must define more than one server to a port on a cluster in order to perform load balancing.
Bind-specific servers: If the Dispatcher component is load balancing to bind-specific servers, then the servers must be configured to bind to the cluster address. Since the Dispatcher forwards packets without changing the destination IP address, when the packets reach the server, the packets will still contain the cluster address as the destination. If a server has been configured to bind to an IP address other than the cluster address, then the server will be unable to accept packets/requests destined for the cluster.
Multiple address collocation: In a collocated configuration, the address of the collocated server machine does not have to be identical to the nonforwarding address (NFA). You can use another address if your machine has been defined with multiple IP addresses. For the Dispatcher component, the collocated server machine must be defined as collocated using the ndcontrol server command. For more information on collocated servers, see Using collocated servers.
For more information on ndcontrol server command syntax, see ndcontrol server -- configure servers.
The manager function improves load balancing. To start the manager, enter the ndcontrol manager start command, edit the sample configuration file, or use the GUI.
The advisors give the manager more information about the ability of the load-balanced server machines to respond to requests. An advisor is specific to a protocol. For example, to start the HTTP advisor, issue the following command:
cbrcontrol advisor start http portFor a list of advisors along with their default ports, see Appendix B, Command reference for Dispatcher, CBR, and Mailbox Locator. For a description of each advisor, see List of advisors.
If you start advisors, you may modify the proportion of importance given to advisor information being included in the load balancing decisions. To set the cluster proportions, issue the ndcontrol cluster set cluster proportions command. For more information, see Proportion of importance given to status information.
If the server is collocated (Dispatcher resides on the same machine it load balances) or if using nat or cbr forwarding methods, do not perform the following procedures.
When using mac forwarding method, Dispatcher will only work with backend servers that allow the loopback adapter to be configured with an additional IP address, for which the backend server will never respond to ARP (address resolution protocol) requests. Follow the steps in this section to set up the load-balanced server machines.
For the load-balanced server machines to work, you must set (or preferably alias) the loopback device (often called lo0) to the cluster address. When using the mac forwarding method, the Dispatcher component does not change the destination IP address in the TCP/IP packet before forwarding the packet to a TCP server machine. By setting or aliasing the loopback device to the cluster address, the load balanced server machines will accept a packet that was addressed to the cluster address.
If you have an operating system that supports network interface aliasing (such as AIX, Linux, Solaris, or Windows 2000), you should alias the loopback device to the cluster address. The benefit of using an operating system that supports aliases is that you have the ability to configure the load-balanced server machines to serve multiple cluster addresses.
For Linux kernel versions 2.2.14 or higher, issue the following commands prior to the ifconfig command:
echo 1 > /proc/sys/net/ipv4/conf/lo/hidden echo 1 > /proc/sys/net/ipv4/conf/all/hidden
If you have a server with an operating system that does not support aliases, such as HP-UX and OS/2, you must set the loopback device to the cluster address.
Use the command for your operating system as shown in Table 4 to set or alias the loopback device.
Table 4. Commands to alias the loopback device (lo0) for Dispatcher
On some operating systems, a default route may have been created and needs to be removed.
route print
netstat -nr
Windows 2000 Example:
Active Routes: Network Address Netmask Gateway Address Interface Metric 0.0.0.0 0.0.0.0 9.67.128.1 9.67.133.67 1 9.0.0.0 255.0.0.0 9.67.133.158 9.67.133.158 1 9.67.128.0 255.255.248.0 9.67.133.67 9.67.133.67 1 9.67.133.67 255.255.255.255 127.0.0.1 127.0.0.1 1 9.67.133.158 255.255.255.255 127.0.0.1 127.0.0.1 1 9.255.255.255 255.255.255.255 9.67.133.67 9.67.133.67 1 127.0.0.0 255.0.0.0 127.0.0.1 127.0.0.1 1 224.0.0.0 224.0.0.0 9.67.133.158 9.67.133.158 1 224.0.0.0 224.0.0.0 9.67.133.67 9.67.133.67 1 255.255.255.255 255.255.255.255 9.67.133.67 9.67.133.67 1
9.0.0.0 255.0.0.0 9.67.133.158 9.67.133.158 1
You must delete the extra route. Use the command for your operating system shown in Table 5 to delete the extra route.
Example: To delete the extra route as shown in the "Active Routes" example table for Step 2, enter:
route delete 9.0.0.0 9.67.133.158
Table 5. Commands to delete any extra route for Dispatcher
Using the example shown in Figure 15, and setting up a server machine that is running AIX, the command would be:
route delete -net 204.0.0.0 204.67.172.72
To verify if a backend server is properly configured, perform the following steps from a different machine on the same subnet when the Network Dispatcher is not running and cluster is unconfigured:
arp -d cluster
ping cluster
There should be no response. If there is a response to the ping, ensure that you did not ifconfig the cluster address to the interface. Ensure that no machine has a published arp entry to the cluster address.
For Linux kernel versions 2.2.14 or higher ensure that there is a "1" in /proc/sys/net/ipv4/conf/lo/hidden and /proc/sys/net/ipv4/conf/all/hidden.
arp -a
In the output from the command, you should see the MAC address of your server. Issue the command:
arp -s cluster server_mac_address
arp -d cluster
For Linux servers only, a specific patch (depending on the Linux kernel version) is required to alias the loopback device.
The patch ensures that an ARP response is sent only from a network adapter port that has the IP address requested in the ARP request. Without this patch, Linux will issue ARP responses on the network for loopback aliases. The patch also corrects an ARP race condition when multiple network adapter ports with different IP addresses are on the same physical network.
You must install the patch under the following conditions.
If you are using the 2.2.12 or 2.2.13 kernel on a backend server.
Notes:
The kernel patch is not required for all configurations. You must install a patch for Linux kernel 2.4.x versions under the following conditions:
You can download this patch from: http://oss.software.ibm.com/developerworks/opensource/cvs/naslib.
Select CVS Tree in the Download list.
To apply the patch:
cd /usr/src/linux-2.4/net/ipv4 patch -p0 -l < arp.c.2.4.0.patch
make dep;make clean;make bzImage;make modules;make modules_install cd arch/i386/boot cat bzImage > /boot/vmlinuz-2.4.2-2-arppatch cd /usr/src/linux-2.4 cp System.map /boot/System.map-2.4.2-2-arppatch cd /etc
A patch for Linux kernel versions 2.2.12 and 2.2.13 must be installed on any server box using the MAC forwarding method. You can download this patch from:http://www.ibm.com/developer/linux.
To apply the patch:
patch -p0 < patchfile
echo 1 > /proc/sys/net/ipv4/conf/lo/arp_invisible
This command will only last until the machine is rebooted. Once rebooted it will be necessary to follow this and the subsequent steps again.
ifconfig lo:1 cluster netmask 255.255.255.255 up
This chapter describes what the network planner should consider before installing and configuring the CBR component with Caching Proxy.
This chapter includes the following sections:
The CBR component allows you to load balance HTTP and SSL traffic using Caching Proxy to proxy the request
CBR is very similar to Dispatcher in its component structure. CBR consists of the following functions:
Using the manager is optional. However, if the manager is not used, load balancing will be performed using weighted round-robin scheduling based on the current server weights, and advisors will not be available.
The three key functions of CBR (executor, manager, and advisors) interact to balance and dispatch the incoming requests between servers. Along with load balancing requests, the executor monitors the number of new connections and active connections and supplies this information to the manager.
The CBR component gives you the ability to specify a set of servers that will handle a request based on regular expression matching the content of the request. CBR allows you to partition your site so that different content or application services can be served by different sets of servers. This partitioning will be transparent to clients accessing your site. Because CBR allows you to specify multiple servers for each type of request, the requests can be load balanced for optimal client response. By allowing multiple servers to be assigned to each type of content, you are protected if one workstation or server fails. CBR will recognize the failure and continue to load balance client requests to the other servers in the set.
One way to divide your site would be to assign some servers to handle only cgi requests, and another set of servers to handle all other requests. This would stop compute intensive cgi scripts from slowing down the servers for normal html traffic, allowing clients to get better overall response time. Using this scheme, you could also assign more powerful workstations for normal requests. This would give clients better response time without the expense of upgrading all your servers. You could also assign more powerful workstations for cgi requests.
Another possibility for partitioning your site could be to direct clients who are accessing pages requiring registration to one set of servers, and all other requests to a second set of servers. This would keep casual browsers of your site from tying up resources that could be used by clients who have committed to your registration. It would also allow you to use more powerful workstations to service those clients who have registered.
You could of course combine the methods above for even more flexibility, and improved service.
Caching Proxy communicates with CBR through its plugin interface. Caching Proxy, must be installed on the same machine. Multiple instances of Caching Proxy running on the same machine can now communicate with CBR simultaneously. In earlier releases, only one instance of Caching Proxy could communicate with CBR.
CBR along with Caching Proxy examines HTTP requests using specified rule types. When running, Caching Proxy accepts client requests and queries the CBR component for the best server. Upon this query, CBR matches the request to a set of prioritized rules. When a rule is matched, an appropriate server is chosen from a preconfigured server set. Finally, CBR informs Caching Proxy which server was chosen and the request gets proxied there.
Once you have defined a cluster to be load balanced, you must make sure that all requests to that cluster have a rule that will choose a server. If no rule is found that matches a particular request, the client will receive an error page from Caching Proxy. The easiest way to ensure that all requests will match some rule, is to create an always true rule at a very high priority number. Make sure that the servers used by this rule can handle all the requests not explicitly handled by the rules that have a lower-numbered priority. (Note: The lower-numbered priority rules are evaluated first.)
CBR with Caching Proxy can receive SSL transmission from the client to the proxy (client-to-proxy side) as well as support transmission from the proxy to an SSL server (proxy-to-server side). By defining an SSL port on a server in the CBR configuration to receive the SSL request from the client, you have the ability to maintain a fully secure site, using CBR to load balance across secure (SSL) servers.
A configuration statement needs to be added to the ibmproxy.conf file for IBM Caching Proxy to enable SSL encryption on the proxy-to-server side. The format must be:
proxy uri_pattern url_pattern address
where uri_pattern is a pattern to match (for example: /secure/*), url_pattern is a replacement URL (for example: https://clusterA/secure/*), and address is the cluster address (for example: clusterA).
CBR with Caching Proxy can also receive SSL transmission from the client and then decrypt the SSL request before proxying the request to an HTTP server. For CBR to support client-to-proxy in SSL and proxy-to-server in HTTP, there is an optional keyword mapport on the cbrcontrol server command. Use this keyword when you need to indicate that the port on the server is different from the incoming port from the client. The following is an example of adding a port using the mapport keyword, where the client's port is 443 (SSL) and the server's port is 80 (HTTP):
cbrcontrol server add cluster:443 mapport 80
The port number for mapport can be any positive integer value. The default is the port number value of the incoming port from the client.
Since CBR must be able to advise on an HTTP request for a server configured on port 443 (SSL), a special advisor ssl2http is provided. This advisor starts on port 443 (the incoming port from the client) and advises on the server(s) configured for that port. If there are two clusters configured and each cluster has port 443 and servers configured with a different mapport, then a single instance of the advisor can open the appropriate port accordingly. The following is an example of this configuration:
Executor Cluster1 Port:443 Server1 mapport 80 Server2 mapport 8080 Cluster2 Port:443 Server3 mapport 80 Server4 mapport 8080 Manager Advisor ssl2http 443
Before following the steps in this chapter, see Planning for the Content Based Routing component. This chapter explains how to create a basic configuration for the CBR component of Network Dispatcher.
Table 6. Configuration tasks for the CBR component
Task | Description | Related information |
---|---|---|
Set up the CBR machine. | Finding out about the requirements. | Setting up the CBR machine |
Set up machines to be load-balanced. | Set up your load balancing configuration. | Step 7. Define load balanced server machines |
To create a basic configuration for the CBR component of Network Dispatcher, there are four basic methods:
To use CBR, Caching Proxy must be installed.
This is the most direct means of configuring CBR. The command parameter values must be entered in English characters. The only exceptions are host names (used, for example, in cluster and server commands) and file names.
To start CBR from the command line:
You can enter an abbreviated version of the cbrcontrol command parameters. You only need to enter the unique letters of the parameters. For example, to get help on the file save command, you can type cbrcontrol he f instead of cbrcontrol help file.
To start up the command line interface: issue cbrcontrol to receive a cbrcontrol command prompt.
To end the command line interface: issue exit or quit.
Notes:
( ) right and left parentheses
& ampersand
| vertical bar
! exclamation point
* asterisk
The operating system's shell may interpret these as special characters and convert them to alternate text before cbrcontrol evaluates them.
The special characters in the above list are optional characters on the cbrcontrol rule add command, and are used when specifying a pattern for a content rule. For example, the following command might be valid only when using the cbrcontrol>> prompt.
rule add 10.1.203.4:80:cbr_prod_rule_ek type content pattern client=181.0.153.222&uri=http://10.1.203.4/nipoek/*
For this same command to work at the operating system's prompt, double quotation marks (" ") must be placed around the pattern as follows:
cbrcontrol rule add 10.1.203.4:80:cbr_prod_rule_ek type content pattern "client=181.0.153.222&uri=http://10.1.203.4/nipoek/*"
If the quotation marks are not used, some of the pattern might be truncated when the rule is saved in CBR. Note that quotation marks are not supported when using the cbrcontrol>> command prompt.
The commands for configuring CBR can be entered into a configuration script file and executed together.
cbrcontrol file appendload myscript
cbrcontrol file newload myscript
For an example of the graphical user interface (GUI), see Figure 2.
To start the GUI, follow these steps
In order to configure the CBR component from the GUI, you must first select Content Based Routing in the tree structure. You can start the manager once you connect to a Host. You can also create clusters containing ports and servers, and start advisors for the manager.
The GUI can be used to do anything that you would do with the cbrcontrol command. For example, to define a cluster using the command line, you would enter cbrcontrol cluster add cluster command. To define a cluster from the GUI, right-click Executor, then in the pop-up menu, left-click Add Cluster. Enter the cluster address in the pop-up window, then click OK.
Pre-existing CBR configuration files can be loaded using the Load New Configuration (for completely replacing the current configuration) and Append to Current Configuration (for updating the current configuration) options presented in the Host pop-up menu. You should save your CBR configuration to a file periodically using the Save Configuration File As option also presented in the Host pop-up menu. The File menu located at the top of the GUI will allow you to save your current host connections to a file or restore connections in existing files across all Network Dispatcher components.
You can access Help by clicking the question mark icon in the upper right hand corner of the Network Dispatcher window.
For more information about using the GUI, see General Instructions for using the GUI.
If you are using the configuration wizard, follow these steps:
Launch the wizard from the command prompt by issuing the cbrwizard. Or, select the Configuration Wizard from the CBR component menu as presented in the GUI.
For AIX, Linux, or Solaris: To start Caching Proxy, enter ibmproxy
For Windows 2000: To start Caching Proxy, go to the Services panel: Start-> Settings-> Control Panel -> Administrative Tools -> Services
The CBR wizard guides you step-by-step through the process of creating a basic configuration for the CBR component. It asks you questions about your network and guides you as you setup a cluster that enables CBR to load balance traffic between a group of servers.
With the CBR configuration wizard, you will see the following panels:
Before setting up the CBR machine, you must be the root user (for AIX, Linux, or Solaris) or the Administrator on Windows 2000.
You will need one IP address for each cluster of servers that will be set up. A cluster address is an address that is associated with a host name (such as www.company.com). This IP address is used by a client to connect to the servers in a cluster. Specifically, this address is found in the URL request from the client. All requests made to the same cluster address are load balanced by CBR.
For Solaris only: Before using the CBR component, the system defaults for IPCs (Inter-process Communication) must be modified. The maximum size of a shared memory segment and the number of semaphore identifiers need to be increased. To tune your system to support CBR, edit the /etc/system file on your system to add the following statements and then reboot:
set shmsys:shminfo_shmmax=0x02000000 set semsys:seminfo_semmap=750 set semsys:seminfo_semmni=30 set semsys:seminfo_semmns=750 set semsys:seminfo_semmnu=30 set semsys:seminfo_semume=30
If you do not increase the shared memory segment to the values shown above, cbrcontrol executor start command will fail.
To use CBR, Caching Proxy must be installed.
You must make the following modifications to the Caching Proxy configuration file (ibmproxy.conf):
Change the incoming URL directive CacheByIncomingUrl to specify "on".
There are four entries that must be edited for the CBR Plug-in:
The specific additions to the configuration file for AIX, Linux, Solaris, and Windows 2000 follow.
Figure 16. CBR configuration file for AIX
ServerInit /usr/lpp/nd/servers/lib/libndcbr.so:ndServerInit PreExit /usr/lpp/nd/servers/lib/libndcbr.so:ndPreExit PostExit /usr/lpp/nd/servers/lib/libndcbr.so:ndPostExit ServerTerm /usr/lpp/nd/servers/lib/libndcbr.so:ndServerTerm
Figure 17. CBR configuration file for Linux
ServerInit /opt/nd/servers/lib/libndcbr.so:ndServerInit PreExit /opt/nd/servers/lib/libndcbr.so:ndPreExit PostExit /opt/nd/servers/lib/libndcbr.so:ndPostExit ServerTerm /opt/nd/servers/lib/libndcbr.so:ndServerTerm
Figure 18. CBR configuration file for Solaris
ServerInit /opt/nd/servers/lib/libndcbr.so:ndServerInit PreExit /opt/nd/servers/lib/libndcbr.so:ndPreExit PostExit /opt/nd/servers/lib/libndcbr.so:ndPostExit ServerTerm /opt/nd/servers/lib/libndcbr.so:ndServerTerm
Figure 19. CBR configuration file for Windows 2000
Common install directory path:
ServerInit c:\Progra~1\IBM\edge\nd\servers\lib\libndcbr.dll:ndServerInit PreExit c:\Progra~1\IBM\edge\nd\servers\lib\libndcbr.dll:ndPreExit PostExit c:\Progra~1\IBM\edge\nd\servers\lib\libndcbr.dll:ndPostExit ServerTerm c:\Progra~1\IBM\edge\nd\servers\lib\libndcbr.dll:ndServerTerm
Native install directory path:
ServerInit c:\Progra~1\IBM\nd\servers\lib\libndcbr.dll:ndServerInit PreExit c:\Progra~1\IBM\nd\servers\lib\libndcbr.dll:ndPreExit PostExit c:\Progra~1\IBM\nd\servers\lib\libndcbr.dll:ndPostExit ServerTerm c:\Progra~1\IBM\nd\servers\lib\libndcbr.dll:ndServerTerm
To start the CBR server function, type cbrserver on the command line.
A default configuration file (default.cfg) gets automatically loaded when starting cbrserver. If you decide to save the CBR configuration in default.cfg, then everything saved in this file will be automatically loaded next time cbrserver gets started.
To start the executor function, enter the cbrcontrol executor start command. You may also change various executor settings at this time. See ndcontrol executor -- control the executor.
CBR will balance the requests sent for the cluster address to the corresponding servers configured on the ports for that cluster.
The cluster address is either a symbolic name or a dotted decimal address. This address will be located in the host portion of the URL.
To define a cluster, issue the following command:
cbrcontrol cluster add cluster
To set cluster options, issue the following command:
cbrcontrol cluster set cluster option value
For more information, see Appendix B, Command reference for Dispatcher, CBR, and Mailbox Locator.
If you are running Caching Proxy configured as a reverse proxy, when load balancing for multiple Web sites, you must add the cluster address for each Web site to at least one of the network interface cards of the Network Dispatcher box. Otherwise, this step can be omitted.
For AIX, Linux, or Solaris: To add the cluster address to
the network interface, use the ifconfig command. Use the command for
your operating system as shown in Table 7.
Table 7. Commands to alias the NIC
AIX | ifconfig interface_name alias cluster_address netmask netmask |
Linux | ifconfig interface_name cluster_address netmask netmask up |
Solaris 7 | ifconfig interface_name cluster_address netmask netmask up |
Solaris 8 | ifconfig addif interface_name cluster_address netmask netmask up |
For Windows: To add the cluster address to the network interface, do the following:
The port number is the port that the server applications are listening on. For CBR with Caching Proxy running HTTP traffic, this is typically port 80.
To define a port to the cluster you defined in the previous step, issue the following:
cbrcontrol port add cluster:port
To set port options, issue the following:
cbrcontrol port set cluster:port option value
For more information, see Appendix B, Command reference for Dispatcher, CBR, and Mailbox Locator.
The server machines are the machines running the applications that you want load balanced. The server is the symbolic name or dotted decimal address of the server machine. To define a server on the cluster and port, issue the following command:
cbrcontrol server add cluster:port:server
You must define more than one server per port on a cluster in order to perform load balancing.
This is the key step in configuring CBR w/Caching Proxy. A rule defines how a URL request will be distinguished and sent to one of the appropriate set of servers. The special rule type used by CBR is called a content rule. To define a content rule, issue the following command:
cbrcontrol rule add cluster:port:rule type content pattern=pattern
The value pattern is the regular expression that will be compared to the URL in each client request. For more information on how to configure the pattern, see Appendix C, Content rule (pattern) syntax.
Some other rule types defined in Dispatcher can also be used in CBR. For more information, see Configure rules-based load balancing.
When a rule is matched by a client request, the rule's set of servers is queried for which server is best. The rule's server set is a subset of the servers defined in the port. To add servers to a rule's server set, issue the following command:
cbrcontrol rule useserver cluster:port:rule server
The manager function improves load balancing. To start the manager, issue the following command:
cbrcontrol manager start
The advisors give the manager more information about the ability of the load balanced server machines to respond to requests. An advisor is specific to a protocol. For example, to start the HTTP advisor, issue the following command:
cbrcontrol advisor start http port
If you start advisors, you may modify the proportion of importance given to advisor information being included in the load balancing decisions. To set the cluster proportions, issue the cbrcontrol cluster set cluster proportions command. For more information, see Proportion of importance given to status information.
/usr/lpp/nd/servers/lib
/opt/nd/servers/lib
Common install directory path:
c:\Program Files\IBM\edge\nd\servers\lib
Native install directory path:
c:\Program Files\IBM\nd\servers\lib
In the new environment, start Caching Proxy: From the command prompt, issue ibmproxy
To configure CBR follow these steps:
This chapter describes what the network planner should consider before installing and configuring the Mailbox Locator component.
This chapter includes the following sections:
The Mailbox Locator component allows you to proxy IMAP and POP3 traffic based on userID and password of the client request.
Mailbox Locator is very similar to Dispatcher in its component structure. Mailbox Locator consists of the following functions:
Using the manager is optional. However, if the manager is not used, load balancing will be performed using weighted round-robin scheduling based on the current server weights, and advisors will not be available.
The three key functions of Mailbox Locator (executor, manager, and advisors) interact to balance and dispatch the incoming requests between servers. Along with load balancing requests, the executor monitors the number of new connections and active connections and supplies this information to the manager.
To start Mailbox Locator, issue the mlserver command from the command prompt.
Mailbox Locator can provide a single point of presence for many IMAP or POP3 servers. Each server can have a subset of all mailboxes serviced by the point of presence. For IMAP and POP3, Mailbox Locator is a proxy that chooses an appropriate server based on userID and password provided by the client.
An example of a method for distributing requests based on client userID is the following. If you have two (or more) POP3 servers, you can choose to divide the mailboxes alphabetically by userID. Client requests with userID's beginning with letters A-I can be distributed to server 1. Client requests with userID's beginning with letters J-R can be distributed on server 2, and so on.
You can also choose to have each mailbox represented on more than one server. In that case, the content of each mailbox must be available to all servers with that mailbox. In the event of a server failure, another server can still access the mailbox.
In order to have only one address representing multiple POP3 mail servers, Mailbox Locator can be configured with a single cluster address that becomes the POP3 mail server address for all clients. The commands to configure this are the following:
mlcontrol cluster add pop3MailServer mlcontrol port add pop3MailServer:110 protocol pop3 mlcontrol server add pop3MailServer:110:pop3Server1+pop3Server2+pop3Server3
In this example, pop3MailServer represents the cluster address. Port 110 with proxy protocol POP3 is added to the pop3MailServer. Pop3Server1, pop3Server2, and pop3Server3 represent POP3 mail servers which are added to the port. With this configuration, you can configure your mail clients' incoming POP3 requests with the pop3MailServer cluster address.
When a POP3 or IMAP request arrives at the proxy, the proxy attempts to contact all the configured servers for the port using the client's userID and password. The client's request is directed to the first server that responds. You should use the sticky/affinity feature in conjunction with the Mailbox Locator for IMAP or POP3 servers. The affinity feature allows subsequent requests from the same client's userID to be directed to the same server. Set stickytime for the port to a value greater than zero to enable this affinity feature. For more information on the affinity feature, see How affinity feature for Network Dispatcher works.
The inactivity autologout timer for POP3 and IMAP protocols is a minimum of 10 minutes and 30 minutes respectively. This timeout is the number of seconds during which there can be no activity on a connection before that connection is removed. To optimize performance, Mailbox Locator overrides the inactivity timeout value to 60 seconds. In order to change the inactivity timeout, change the staletimeout value on the mlcontrol port command. For information on configuring this command, see ndcontrol port -- configure ports.
Before following the steps in this chapter, see Planning for the Mailbox Locator component. This chapter explains how to create a basic configuration for the Mailbox Locator component of Network Dispatcher.
Table 8. Configuration tasks for the Mailbox Locator component
Task | Description | Related information |
---|---|---|
Set up the Mailbox Locator machine. | Finding out about the requirements. | Setting up the Mailbox Locator machine |
Set up machines to be load-balanced. | Set up your load balancing configuration. | Step 4. Define load balanced server machines |
To create a basic configuration for the Mailbox Locator component of Network Dispatcher, there are four basic methods:
This is the most direct means of configuring Mailbox Locator. The command parameter values must be entered in English characters. The only exceptions are host names (used, for example, in cluster and server commands) and file names.
To start Mailbox Locator from the command line:
You can enter a minimized version of the mlcontrol command parameters. You only need to enter the unique letters of the parameters. For example, to get help on the file save command, you can type mlcontrol he f instead of mlcontrol help file.
To start up the command line interface: issue mlcontrol to receive an mlcontrol command prompt.
To end the command line interface: issue exit or quit.
The commands for configuring Mailbox Locator can be entered into a configuration script file and executed together.
mlcontrol file appendload myscript
mlcontrol file newload myscript
For an example of the GUI, see Figure 2.
To start the GUI, follow these steps
In order to configure the Mailbox Locator component from the GUI, you must first select Mailbox Locator in the tree structure. You can start the manager once you connect to a Host. You can also create clusters containing ports and servers, and start advisors for the manager.
The GUI can be used to do anything that you would do with the mlcontrol command. For example, to define a cluster using the command line, you would enter mlcontrol cluster add cluster command. To define a cluster from the GUI, right-click Executor, then in the pop-up menu, left-click Add Cluster. Enter the cluster address in the pop-up window, then click OK.
Pre-existing Mailbox Locator configuration files can be loaded using the Load New Configuration (for completely replacing the current configuration) and Append to Current Configuration (for updating the current configuration) options presented in the Host pop-up menu. You should save your Mailbox Locator configuration to a file periodically using the Save Configuration File As option also presented in the Host pop-up menu. The File menu located at the top of the GUI will allow you to save your current host connections to a file or restore connections in existing files across all Network Dispatcher components.
You can access Help by clicking the question mark icon in the upper right hand corner of the Network Dispatcher window.
For more information about using the GUI, see General Instructions for using the GUI.
If you are using the configuration wizard, follow these steps:
You can launch this wizard from the command prompt by issuing the mlwizard. Or, select the Configuration Wizard from the Mailbox Locator component menu as presented in the GUI.
The Mailbox Locator wizard guides you step-by-step through the process of creating a basic configuration for the Mailbox Locator component. It asks you questions about your network and guides you as you setup a cluster that enables Mailbox Locator to load balance traffic between a group of servers.
With the Mailbox Locator configuration wizard, you will see the following panels:
Before setting up the Mailbox Locator machine, you must be the root user (for AIX, Linux, or Solaris) or the Administrator on Windows 2000.
You will need one IP address for each cluster of servers that will be set up. A cluster address is an address that is associated with a host name (such as www.yourcompany.com). This IP address is used by a client to connect to the servers in a cluster. All requests made to the same cluster address are load balanced by Mailbox Locator.
To start the server function, type mlserver on the command line.
Mailbox Locator will balance the requests sent for the cluster address to the corresponding servers configured on the ports for that cluster.
The cluster address is either the symbolic name or a dotted decimal address.
To define a cluster, issue the following command:
mlcontrol cluster add cluster
To set cluster options, issue the following command:
mlcontrol cluster set cluster option value
For more information, see Appendix B, Command reference for Dispatcher, CBR, and Mailbox Locator.
The port number is the port that the server applications are listening on. For IMAP traffic, this is typically port 143. And, for POP3 traffic, this is typically port 110.
To define a port to the cluster you defined in the previous step, issue the following:
mlcontrol port add cluster:port protocol [pop3|imap]
To set port options, issue the following:
mlcontrol port set cluster:port option value
For more information, see Appendix B, Command reference for Dispatcher, CBR, and Mailbox Locator.
The mail servers are the machines running the applications that you want load balanced. The server is the symbolic name or dotted decimal address of the server machine. To define a server on the cluster and port from step 3, issue the following command:
mlcontrol server add cluster:port:server
You must define more than one server per port on a cluster in order to perform load balancing.
The manager function improves load balancing. To start the manager, issue the following command:
mlcontrol manager start
The advisors give the manager more information about the ability of the load balanced server machines to respond to requests. An advisor is specific to a protocol. The Network Dispatcher supplies IMAP and POP3 advisors. For example, to start the IMAP advisor, issue the following command:
mlcontrol advisor start imap portFor a list of advisors along with their default ports, see Appendix B, Command reference for Dispatcher, CBR, and Mailbox Locator. For a description of each advisor, see List of advisors.
If you start advisors, you may modify the proportion of importance given to advisor information being included in the load balancing decisions. To set the cluster proportions, issue the mlcontrol cluster set cluster proportions command. For more information, see Proportion of importance given to status information.
This chapter describes what the network planner should consider before installing and configuring the Site Selector component.
This chapter includes the following sections:
Site Selector works in conjunction with a domain name server to load balance among a group of servers using measurements and weights that are gathered. You can create a site configuration to let you load balance traffic among a group of servers based on the domain name used for a client's request.
Figure 20. Example of a DNS environment
When setting up a subdomain for Site Selector within your DNS environment, Site Selector should have authority over its own subdomain. For example (see Figure 20), your company has been assigned authority over company.com domain. Within the company, there are several subdomains. Site Selector would have authority for siteload.company.com, while the DNS server(s) would still maintain authority for atlanta.company.com and boston.company.com.
In order for company's name server to recognize Site Selector as having authority for the siteload subdomain, a name server entry will need to be added to its named data file. For example, on AIX, a name server entry would look like the following:
siteload.company.com. IN NS siteselector.company.com.
Where siteselector.company.com is the hostname of the Site Selector machine. Equivalent entries would need to be made in any other named database files for use by DNS servers.
A client submits a request for resolution of a domain name to a name server within its network. Name server forwards the request to the Site Selector machine. Site Selector then resolves the domain name to the IP address of one of the servers that has been configured under the site name. Site Selector returns the IP address of the selected server to the name server. Name server returns the IP address to the client. (Site Selector acts as a non-recursive (leaf node) name server, and it will return an error if it does not resolve the domain name request.)
Refer to Figure 11 which illustrates a site in which Site Selector is used in conjunction with a DNS system to load balance across local and remote servers.
Site Selector consists of the following functions:
Using the manager is optional. However, if the manager is not used, load balancing will be performed using weighted round-robin scheduling based on the current server weights, and advisors will not be available.
With Metric Server, Site Selector can monitor the level of activity on a server, detect when a server is the least heavily loaded, and detect a failed server. The load is a measure of how hard the server is working. The system Site Selector administrator controls the type of measurement used to measure the load. You can configure Site Selector to suit your environment, taking into account such factors as frequency of access, the total number of users, and types of access (for example, short queries, long-running queries, or CPU-intensive loads).
Load balancing is based on server weights. For Site Selector, there are four proportions which the manager uses to determine weights:
CPU and memory values are all supplied by Metric Server. Consequently, use of Metric Server is recommended with the Site Selector component.
See Metric Server for more information.
The four key functions of Site Selector (name server, manager, Metric Server, and advisors) interact to balance and resolve the incoming requests between servers.
Using DNS-based load balancing requires that caching of name resolutions be disabled. The TTL (time to live) value determines the effectiveness of DNS-based load balancing. TTL determines how long another nameserver will cache the resolved response. Small TTL values allow for subtle changes in the server or network load to be realized more quickly. However, disabling caching requires that clients contact the authoritative name server for every name resolution request, thus potentially increasing the client latency. When choosing a TTL value, careful consideration should be given to the impact that disabled-caching has on an environment. Also be aware that DNS-based load balancing is potentially limited by client-side caching of name resolutions.
TTL can be configured using the sscontrol sitename [add | set] command. See sscontrol sitename -- configure a sitename, for more information.
Network proximity is the calculation of each server's nearness to the requesting client. To determine network proximity, the Metric Server agent (which must reside on each load-balanced server) sends a ping to the client IP address and returns the response time to Site Selector. Site Selector uses the proximity response in the load-balancing decision. Site Selector combines the network proximity response value with the weight from the manager to create a combined final weight value for the server.
Use of the network proximity feature with Site Selector is optional.
The Site Selector provides the following network proximity options that can be set per site name:
If set to yes, the Metric Server pings the client to obtain the proximity response time. Name server waits for all Metric Servers to respond or for a time-out to occur. Then, for each server, the name server combines the proximity response time with the weight the manager calculated to create a "combined weight" value for each server. Site Selector will supply the client with the server IP address with the best combined weight. (It is expected that most client name servers have a 5 second time-out. Site Selector tries to respond before that time-out is exceeded.)
If set to no, a name resolution will be provided to the client based on the current manager weights. Then, the Metric Server pings the client to obtain the proximity response time. The name server caches the response time it receives from the Metric Server. When the client returns for a second request, the name server combines the current manager weight with the cached ping response value for each server to obtain the server with the best "combined weight." Site Selector returns this server's IP address to the client for its second request.
Network proximity options can be set on the sscontrol sitename [add | set] command. See Appendix D, Command reference for Site Selector for more information.
Before following the steps in this chapter, see Planning for the Site Selector component. This chapter explains how to create a basic configuration for the Site Selector component of Network Dispatcher.
Table 9. Configuration tasks for the Site Selector component
Task | Description | Related information |
---|---|---|
Set up the Site Selector machine. | Finding out about the requirements. | Setting up the Site Selector machine |
Set up machines to be load-balanced. | Set up your load balancing configuration. | Step 4. Define load balanced server machines |
To create a basic configuration for the Site Selector component of Network Dispatcher, there are four basic methods of configuring the Site Selector component:
This is the most direct means of configuring Site Selector. The command parameter values must be entered in English characters. The only exceptions are host names (used, for example, in site name and server commands) and file names.
To start Site Selector from the command line:
You can enter a minimized version of the sscontrol command parameters. You only need to enter the unique letters of the parameters. For example, to get help on the file save command, you can type sscontrol he f instead of sscontrol help file.
To start up the command line interface: issue sscontrol to receive an sscontrol command prompt.
To end the command line interface: issue exit or quit.
The commands for configuring Site Selector can be entered into a configuration script file and executed together.
sscontrol file appendload myscript
sscontrol file newload myscript
For an example of the GUI, see Figure 2.
To start the GUI, follow these steps
In order to configure the Site Selector component from the GUI, you must first select Site Selector in the tree structure. You can start the manager once you connect to a Host. You can also create site names containing ports and servers, and start advisors for the manager.
The GUI can be used to do anything that you would do with the sscontrol command. For example, to define a site name using the command line, you would enter sscontrol sitename add sitename command. To define a site name from the GUI, right-click Name Server, then in the pop-up menu, left-click Add Site Name. Enter the site name in the pop-up window, then click OK.
Pre-existing Site Selector configuration files can be loaded using the Load New Configuration (for completely replacing the current configuration) and Append to Current Configuration (for updating the current configuration) options presented in the Host pop-up menu. You should save your Site Selector configuration to a file periodically using the Save Configuration File As option also presented in the Host pop-up menu. The File menu located at the top of the GUI will allow you to save your current host connections to a file or restore connections in existing files across all Network Dispatcher components.
You can access Help by clicking the question mark icon in the upper right hand corner of the Network Dispatcher window.
For more information about using the GUI, see General Instructions for using the GUI.
If you are using the configuration wizard, follow these steps:
You can launch this wizard from the command prompt by issuing the sswizard. Or, select the Configuration Wizard from the Site Selector component menu as presented in the GUI.
The Site Selector wizard guides you step-by-step through the process of creating a basic configuration for the Site Selector component. It asks you questions about your network and guides you as you setup a site name that enables Site Selector to load balance traffic between a group of servers.
With the Site Selector configuration wizard, you will see the following panels:
Before setting up the Site Selector machine, you must be the root user (for AIX, Linux, or Solaris) or the Administrator on Windows 2000.
You will need an unresolvable DNS hostname to use as a site name for a group of servers that you set up. The site name is the name that the clients use to access your site (such as www.yourcompany.com). Site Selector will load-balance traffic for this site name among the group of servers using DNS.
AIX, Linux, and Solaris: To start the server function, enter ssserver.
To start the Name Server, enter the sscontrol nameserver start command.
Optionally, start the Name Server using the bindaddress keyword to bind only to the specified address.
Site Selector will balance the requests sent for the site name to the corresponding servers configured to it.
The site name is an unresolvable host name that the client will request. The site name must be a fully qualified domain name (e.g. www.dnsdownload.com). When a client requests this site name, one of the server IP addresses associated with the site name will be returned.
To define a site name, issue the following command:
sscontrol sitename add sitename
To set site name options, issue the following command:
sscontrol sitename set sitename option value
For more information, see Appendix D, Command reference for Site Selector.
The server machines are the machines running the applications that you want load balanced. The server is the symbolic name or dotted decimal address of the server machine. To define a server on the site name from step 3, issue the following command:
sscontrol server add sitename:server
You must define more than one server under a site name in order to perform load balancing.
The manager function improves load balancing. Prior to starting the manager function, ensure that the metric server is installed in all the load-balanced machines.
To start the manager, issue the following command:
sscontrol manager start
The advisors give the manager more information about the ability of the load balanced server machines to respond to requests. An advisor is specific to a protocol. The Network Dispatcher supplies many advisors. For example, to start the HTTP advisor for a specific site name, issue the following command:
sscontrol advisor start http sitename:port
See Metric Server for information on using system metrics and Metric Server.
If you start advisors, you may modify the proportion of importance given to advisor (port) information being included in the load balancing decisions. To set the site name proportions, issue the sscontrol sitename set sitename proportions command. For more information, see Proportion of importance given to status information.
It is recommended to use Metric Server with the Site Selector component. Refer to Metric Server for information on setting up Metric Server on all server machines that Site Selector is load balancing.
This chapter describes what the network planner should consider before installing and configuring the Consultant for Cisco CSS Switches component.
This chapter includes:
The configuration for Cisco Consultant is dependent on the configuration for the Cisco CSS Switch (see Table 10). After you complete planning and configuration for the Cisco CSS Switch, you can configure and use Cisco Consultant. Refer to the Cisco CSS Switch documentation for planning and configuration instructions.
Consultant consists of the following:
The advisors query the servers and analyze results by protocol before calling the manager to set weights as appropriate. Currently, Cisco Consultant provides advisors such as HTTP, FTP, SSL, SMTP, NNTP, IMAP, POP3 (and others). You also have the option of writing your own advisors (see Create custom (customizable) advisors). Using the advisors is recommended, but is optional.
Metric Server provides server load information to Consultant in the form of system-specific metrics, reporting on the health of the servers. The manager queries the Metric Server residing on each of the servers, using the metrics gathered from the agents to assist in assigning weights to the load-balancing process. The results are also placed into the manager report.
The manager collects information from the Cisco CSS Switch, the advisors, and Metric Server. Based on the information the manager receives, it adjusts how the server machines are weighted on each port and gives the Cisco CSS Switch the new weighting for use in its balancing of new connections. When the manager discovers that a server is down, it assigns that server a weight of zero, and the server is suspended. Subsequently, the Cisco CSS Switch stops forwarding traffic to that server.
The advisors monitor each server on the assigned port to determine the server's response time and availability and then give this information to the manager. The advisors also monitor whether a server is up or down.
To properly configure Consultant, your configuration must mirror the Cisco CSS Switch configuration. First, refer to the Cisco Services Switch Getting Started Guide to configure the Cisco CSS Switch. Ensure that the switch is working correctly, then configure Consultant.
The Cisco CSS Switch configuration consists of owners, content rules, and
services that map to a Consultant configuration as follows:
Table 10. Consultant and Cisco CSS Switch configuration terms
Cisco CSS Switch | Consultant |
---|---|
virtual IP address (VIP) of one or more of the owner's content rules | cluster |
port contained in the content rule | port |
service | server |
The Consultant configuration tree consists of:
Figure 21. Example of Consultant configured with 2 clusters, each with 3 ports
In Figure 21:
When configuring the executor, you must configure an address and SNMP
community name, and these must match the corresponding attributes on the Cisco
CSS Switch. See lbccontrol executor -- control the executor for information on configuring the
executor.
Table 11. Example of the Cisco CSS Switch configuration mapped to the Consultant configuration
Cisco CSS Switch configuration | Consultant configuration |
---|---|
username admin superuser snmp community community private read-write |
lbccontrol executor set address 10.10.20.1 lbccontrol executor set communityname community |
content rule1 port 80 balance weightedrr add service server1 add service server2 vip address 9.67.154.35 active |
lbccontrol cluster add 9.67.154.35 lbccontrol port add 9.67.154.35:80 |
content rule 2 protocol tcp port 443 balance weightedrr add service server3 add service server4 vip address 9.67.154.35 active |
lbccontrol port add 9.67.154.35:443 |
service server1 ip address 10.10.20.2 port 80 weight 4 active |
lbccontrol server add 9.67.154.35:80:server1 address 10.10.20.2 |
service server3 ip address 10.10.20.4 port 443 weight 4 active |
lbccontrol server add 9.67.154.35:443:server3 address 10.10.20.4 |
Before following the steps in this chapter, see Planning for the Consultant for Cisco CSS Switches component. This chapter explains how to create a basic configuration for the Consultant for Cisco CSS Switches component of Network Dispatcher.
Before you begin any of the configuration methods in this chapter:
Table 12. Configuration tasks for the Consultant for Cisco CSS Switches component
Task | Description | Related information |
---|---|---|
Set up the Consultant for Cisco CSS Switches machine | Finding out about the requirements | Setting up the Consultant for Cisco CSS Switches machine |
Test your configuration | Confirm that the configuration is working | Testing your configuration |
To create a basic configuration for the Consultant for Cisco CSS Switches component of Network Dispatcher, there are three methods:
This is the most direct means of configuring Cisco Consultant. The procedures in this manual assume use of the command line. The command parameter values must be entered in English characters. The only exceptions are host names (used, for example, in cluster and server commands) and file names.
To start Cisco Consultant from the command line:
You can enter an abbreviated version of the lbccontrol command parameters. You only need to enter the unique letters of the parameters. For example, to get help on the file save command, you can type lbccontrol he f instead of lbccontrol help file.
To start up the command line interface: issue lbccontrol to receive an lbccontrol command prompt.
To end the command line interface: issue exit or quit.
The commands for configuring Consultant for Cisco CSS Switches can be entered into configuration script file and executed together.
lbccontrol file appendload myscript
lbccontrol file newload myscript
For an example of the graphical user interface (GUI), see Figure 2.
To start the GUI, follow these steps
lbcserver.
To configure the Cisco Consultant component from the GUI:
You can use the GUI to do anything that you would do with the lbccontrol command. For example, to define a cluster using the command line, you would enter lbccontrol cluster add cluster command. To define a cluster from the GUI, right-click Executor, then click Add Cluster. Enter the cluster address in the pop-up window, then click OK.
You can use the Load New Configuration (for completely replacing the current configuration) and Append to Current Configuration (for updating the current configuration) options presented in the Host pop-up menu to load pre-existing Cisco Consultant configuration files. Select Save Configuration File As option to periodically save your Cisco Consultant configuration to a file. Click File on the menu bar to save your current host connections to a file or restore connections in existing files across all Network Dispatcher components.
To access Help click the question mark icon in the upper right corner of the Network Dispatcher window.
For more information about using the GUI, see General Instructions for using the GUI.
Before setting up the Consultant for Cisco CSS Switches machine, you must be the root user (for AIX, Linux, or Solaris) or the Administrator on Windows 2000.
Consultant must be able to connect to the Cisco CSS Switch as a Cisco CSS Switch administrator.
When configuring the executor, you must configure the address and SNMP community name must match the corresponding attributes on the Cisco CSS Switch.
For help with commands used in this procedure, see Appendix E, Command reference for Consultant for Cisco CSS Switches.
If the lbcserver is not already running, start it now by running the following as root:
lbcserver
You must configure an address and SNMP community name. These values must match the corresponding attributes on the Cisco CSS Switch.
Cluster is either a resolvable name or the dotted-decimal address. The cluster corresponds to the Cisco CSS Switch's virtual IP address of an owner's content rule.
To define a cluster, type lbccontrol cluster add cluster. To set cluster options, type lbccontrol cluster set.
To define a port, type lbccontrol port add cluster:port. The port corresponds to the port configured in the Cisco CSS Switch content rule for the owner.
Port is the number of the port you are using for that protocol as specified in the owner content rule for the Cisco CSS Switch. See lbccontrol port -- configure ports for more information.
You can configure multiple instances of the same server within any cluster and port. (Remember that the address and SNMP community name must match the corresponding attributes on the Cisco CSS Switch.) When you configure multiple instances of the same server, you can distinguish different application servers that reside on the same physical machine and respond to the same IP address on the same port.
To define a load-balanced server machine, type:
lbccontrol server add cluster:port:server address x.x.x.x | hostname
The server corresponds to the Cisco CSS Switch service name.
You must define more than one server to a port on a cluster to perform load balancing, or traffic will be directed to only one server. See Server Partitioning: logical servers configured to one physical server (IP address).
For more information on lbccontrol server command syntax, see lbccontrol server -- configure servers.
To start the manager, type the lbccontrol manager start command. See lbccontrol manager -- control the manager for more information.
The advisors give the manager more information about the ability of the load-balanced server machines to respond to requests. An advisor is specific to a protocol. For example, to start the HTTP advisor, issue the following command:
lbccontrol advisor start http portFor a list of advisors along with their default ports, see lbccontrol advisor -- control the advisor. For a description of each advisor, see List of advisors.
If you start any advisors, you must change the cluster proportions so that advisor information is included in the load-balancing decisions. Use the lbccontrol cluster proportions command. See Proportion of importance given to status information.
See Metric Server for information on using Metric Server.
Test to see if the configuration is working.
This chapter explains how to configure the load balancing parameters of Network Dispatcher and how to set up Network Dispatcher for advanced functions.
Table 13. Advanced configuration tasks for the Network Dispatcher
Task | Description | Related information |
---|---|---|
Optionally, change load-balancing settings |
You can change the following load-balancing settings:
| Optimizing the load balancing provided by Network Dispatcher |
Use scripts to generate an alert or record server failure when manager marks server(s) down/up | Network Dispatcher provides user exits that trigger scripts that you can customize when the manager marks server(s) down/up | Using scripts to generate an alert or record server failure |
Use advisors and create custom advisors | Describes the advisors and how to write your own custom advisors for reporting on specific statuses of your servers | Advisors Create custom (customizable) advisors |
Use Workload Manager advisor (WLM) | WLM advisor provides system load information to Network Dispatcher | Workload Manager advisor |
Use Metric Server agent | Metric Server provides system load information to Network Dispatcher | Metric Server |
Use Server partitioning | Define logical servers to distribute load based on services provided | Server Partitioning: logical servers configured to one physical server (IP address) |
Use advisor request/response (URL) option | Define a unique client HTTP URL string, specific for a service that you want to query on the machine | HTTP advisor request/response (URL) option |
Collocate Network Dispatcher on a machine that it is load balancing | Set up a collocated Network Dispatcher machine. | Using collocated servers |
Configure wide area Dispatcher support | Set up a remote Dispatcher to load balance across a wide area network. Or, load balance across a wide area network (without a remote Dispatcher) using a server platform that supports GRE. | Configure wide area Dispatcher support |
Configure high availability or mutual high availability | Set up a second Dispatcher machine to provide a backup. | High availability |
Configure rules-based load balancing | Define conditions under which a subset of your servers will be used. | Configure rules-based load balancing |
Use explicit linking | Avoid bypassing the Dispatcher in your links. | Using explicit linking |
Use a private network | Configure the Dispatcher to load balance servers on a private network. | Using a private network configuration |
Use wildcard cluster to combine common server configurations | Addresses that are not explicitly configured will use the wildcard cluster as a way to load balance traffic. | Use wildcard cluster to combine server configurations |
Use wildcard cluster to load balance firewalls | All traffic will be load balanced to firewalls. | Use wildcard cluster to load balance firewalls |
Use wildcard cluster with Caching Proxy for transparent proxy | Allows Dispatcher to be used to enable a transparent proxy. | Use wildcard cluster with Caching Proxy for transparent proxy |
Use wildcard port to direct unconfigured port traffic | Handles traffic that is not configured for any specific port. | Use wildcard port to direct unconfigured port traffic |
Use sticky affinity feature to configure a cluster's port to be sticky | Allows client requests to be directed to the same server. | How affinity feature for Network Dispatcher works |
Use Server Directed Affinity API | Provides an API which allows an external agent to influence the Dispatcher affinity behavior | Server Directed Affinity API to control client-server affinity |
Use cross port affinity to expand the sticky (affinity) feature across ports | Allows client requests received from different ports to be directed to the same server. | Cross port affinity |
Use affinity address mask to designate a common IP subnet address | Allows clients requests received from the same subnet to be directed to the same server. | Affinity address mask |
Use rule affinity override to provide a mechanism for a server to override the port sticky feature | Allows a server to override the stickytime setting on its port. | Rule affinity override |
Use active cookie affinity to load balance servers for CBR | A rule option that allows a session to maintain affinity for a particular server. | Active cookie affinity |
Use passive cookie affinity to load balance servers for Dispatcher's content-based routing and the CBR component | A rule option that allows a session to maintain affinity for a particular server based on the cookie name/cookie value. | Passive cookie affinity |
Use URI affinity to load-balance across Caching Proxy servers with unique content to be cached on each individual server | A rule option that allows a session to maintain affinity for a particular server based on the URI. | URI affinity |
Use "Denial of Service Attack" detection to notify administrators (via an alert) of potential attacks | Dispatcher analyzes incoming requests for a conspicuous amount of half-open TCP connections on servers. | Denial of service attack detection |
Use binary logging to analyze server statistics | Allows server information to be stored in and retrieved from binary files. | Using binary logging to analyze server statistics |
Using Cisco Consultant (additional information) | How Cisco Consultant interacts with Cisco CSS Switch and additional information on configuring weights. | Additional information on advanced Cisco Consultant functions |
The manager function of Network Dispatcher performs load balancing based on the following settings:
You can change these settings to optimize load balancing for your network.
The manager can use some or all of the following external factors in its weighting decisions:
Or --
Cpu: The percentage of CPU in use on each load balanced server machine (input from Metric Server agent). For Site Selector only, this proportion appears in place of the active connection proportion column.
Or --
Memory: The percentage of memory in use (input from Metric Server agent) on each load balanced server. For Site Selector only, this proportion appears in place of the new connection proportion column.
Along with the current weight for each server and some other information required for its calculations, the manager gets the first two values (active and new connections) from the executor. These values are based on information that is generated and stored internally in the executor.
You can change the relative proportion of importance of the four values on a per cluster (or site name) basis. Think of the proportions as percentages; the sum of the relative proportions must equal 100%. The default ratio is 50/50/0/0, which ignores the advisor and system information. In your environment, you may need to try different proportions to find the combination that gives the best performance.
When adding the WLM advisor, if the system metric proportion is zero, then the manager increases this value to 1. Since the sum of the relative proportions must total 100, the highest value is then decreased by 1.
The number of active connections is dependent upon the number of clients as well as the length of time necessary to use the services that are being provided by the load balanced server machines. If the client connections are quick (such as small Web pages served using HTTP GET), then the number of active connections will be fairly low. If the client connections are slower (such as a database query), then the number of active connections will be higher.
You should avoid setting active and new connections proportions values too low. You will disable Network Dispatcher's load balancing and smoothing unless you have these first two values set to at least 20 each.
To set the proportion of importance values use the ndcontrol cluster set cluster proportions command. See ndcontrol cluster -- configure clusters for more information.
Weights are set by the manager function based upon internal counters in the executor, feedback from the advisors, and feedback from a system-monitoring program, such as Metric Server. If you want to set weights manually while running the manager, specify the fixedweight option on the ndcontrol server command. For a description of the fixedweight option, see Manager fixed weights.
Weights are applied to all servers on a port. For any particular port, the requests will be distributed between servers based on their weights relative to each other. For example, if one server is set to a weight of 10, and the other to 5, the server set to 10 should get twice as many requests as the server set to 5.
To specify the maximum weight boundary that any server can have, enter the ndcontrol port set weightbound command. This command affects how much difference there can be between the number of requests each server will get. If you set the maximum weight to 1, then all the servers can have a weight of 1, 0 if quiesced, or -1 if marked down. As you increase this number, the difference in how servers can be weighted is increased. At a maximum weight of 2, one server could get twice as many requests as another. At a maximum weight of 10, one server could get 10 times as many requests as another. The default maximum weight is 20.
If an advisor finds that a server has gone down, it tells the manager, which sets the weight for the server to zero. As a result, the executor will not send any additional connections to that server as long as that weight remains zero. If there were any active connections to that server before the weight changed, they will be left to complete normally.
Without the manager, advisors cannot be run and cannot detect if a server is down. If you choose to run the advisors, but do not want the manager to update the weight you have set for a particular server, use the fixedweight option on the ndcontrol server command. For example:
ndcontrol server set cluster:port:server fixedweight yes
After fixedweight is set to yes, use the ndcontrol server set weight command to set the weight to the value you desire. The server weight value remains fixed while the manager is running until you issue another ndcontrol server command with fixedweight set to no. For more information, see ndcontrol server -- configure servers.
To optimize overall performance, the manager is restricted in how often it can interact with the executor. You can make changes to this interval by entering the ndcontrol manager interval and ndcontrol manager refresh commands.
The manager interval specifies how often the manager will update the server weights that the executor uses in routing connections. If the manager interval is too low, it can mean poor performance as a result of the manager constantly interrupting the executor. If the manager interval is too high, it can mean that the executor's request routing will not be based on accurate, up-to-date information.
For example, to set the manager interval to 1 second, enter the following command:
ndcontrol manager interval 1
The manager refresh cycle specifies how often the manager will ask the executor for status information. The refresh cycle is based on the interval time.
For example, to set the manager refresh cycle to 3, enter the following command:
ndcontrol manager refresh 3
This will cause the manager to wait for 3 intervals before asking the executor for status.
Network Dispatcher provides other methods for you to optimize load balancing for your servers. To work at top speed, updates to the weights for the servers are only made if the weights have changed significantly. Constantly updating the weights when there is little or no change in the server status would create an unnecessary overhead. When the percentage weight change for the total weight for all servers on a port is greater than the sensitivity threshold, the manager updates the weights used by the executor to distribute connections. Consider, for example, that the total weight changes from 100 to 105. The change is 5%. With the default sensitivity threshold of 5, the manager will not update the weights used by the executor, because the percentage change is not above the threshold. If, however, the total weight changes from 100 to 106, the manager will update the weights. To set the manager's sensitivity threshold to a value other than the default (for example, 6), enter the following command:
ndcontrol manager sensitivity 6
In most cases, you will not need to change this value.
The manager calculates the server weights dynamically. As a result, an updated weight can be very different from the previous one. Under most circumstances, this will not be a problem. Occasionally, however, it may cause an oscillating effect in the way the requests are load balanced. For example, one server can end up receiving most of the requests due to a high weight. The manager will see that the server has a high number of active connections and that the server is responding slowly. It will then shift the weight over to the free servers and the same effect will occur there too, creating an inefficient use of resources.
To alleviate this problem, the manager uses a smoothing index. The smoothing index limits the amount that a server's weight can change, effectively smoothing the change in the distribution of requests. A higher smoothing index will cause the server weights to change less drastically. A lower index will cause the server weights to change more drastically. The default value for the smoothing index is 1.5. At 1.5, the server weights can be rather dynamic. An index of 4 or 5 will cause the weights to be more stable. For example, to set the smoothing index to 4, enter the following command:
ndcontrol manager smoothing 4
In most cases, you will not need to change this value.
Network Dispatcher provides user exits that trigger scripts that you can customize. You can create the scripts to perform automated actions, such as alerting an Administrator when servers are marked down by the manager or simply record the event of the failure. Sample scripts, which you can customize, are in the ...nd/servers/samples install directory. In order to run the files, you must move them to the ...nd/servers/bin directory and remove the ".sample" file extension. The following sample scripts are provided:
Advisors are agents within Network Dispatcher. Their purpose is to assess the health and loading of server machines. They do this with a proactive client-like exchange with the servers. Advisors can be considered as lightweight clients of the application servers.
The product provides several protocol-specific advisors for the most popular protocols. However, it does not make sense to use all of the provided advisors with every component of Network Dispatcher. (For instance, you would not use the Telnet advisor with the CBR component.) Network Dispatcher also supports the concept of a "custom advisor" that allows users to write their own advisors.
Limitation for bind-specific server applications on Linux: For Linux, Network Dispatcher does not support the use of advisors when load-balancing servers with bind-specific server applications (including other Network Dispatcher components such as Mailbox Locator or Site Selector) when they are binding to the cluster IP address.
Advisors periodically open a TCP connection with each server and send a request message to the server. The content of the message is specific to the protocol running on the server. For example, the HTTP advisor sends an HTTP "HEAD" request to the server.
Advisors then listen for a response from the server. After getting the response, the advisor makes an assessment of the server. To calculate this "load" value, most advisors measure the time for the server to respond, and then use this value (in milliseconds) as the load.
Advisors then report the load value to the manager function, where it appears in the manager report in the "Port" column. The manager then calculates aggregate weight values from all its sources, per its proportions, and sets these weight values into the executor function. The Executor will then use these weights for load balancing new incoming client connections.
If the advisor determines that a server is alive and well, it will report a positive, non-zero load number to the Manager. If the advisor determines that a server is not active, it will return a special load value of negative one (-1). The Manager and the Executor will not forward any further connections to that server.
You can start an advisor for a particular port across all clusters (group advisor). Or, you can choose to run different advisors on the same port, but on different clusters (cluster/site specific advisor). For example, if you have Network Dispatcher defined with three clusters (clusterA, clusterB, clusterC), each having port 80 you can do the following:
ndcontrol advisor start http clusterA:80This command will start the http advisor on port 80 for clusterA. The http advisor will advise on all servers attached to port 80 for clusterA.
ndcontrol advisor start ADV_custom 80This command will start the ADV_custom advisor on port 80 for clusterB and clusterC. Your custom advisor will advise on all servers attached to port 80 for clusterB and clusterC. (For more information on custom advisors, see Create custom (customizable) advisors.)
Using the above configuration example for the group advisor, you can choose to stop the custom advisor ADV_custom for port 80 on just one of the clusters or for both clusters (clusterB and clusterC).
ndcontrol advisor stop ADV_custom clusterB:80
ndcontrol advisor stop ADV_custom 80
The advisor interval sets how often an advisor asks for status from the servers on the port it is monitoring and then reports the results to the manager. If the advisor interval is too low, it can mean poor performance as a result of the advisor constantly interrupting the servers. If the advisor interval is too high, it can mean that the manager's decisions about weighting will not be based on accurate, up-to-date information.
For example, to set the interval to 3 seconds for the HTTP advisor for port 80, enter the following command:
ndcontrol advisor interval http 80 3
It does not make sense to specify an advisor interval that is smaller than the manager interval. The default advisor interval is seven seconds.
To make sure that out-of-date information is not used by the manager in its load-balancing decisions, the manager will not use information from the advisor whose time stamp is older than the time set in the advisor report timeout. The advisor report timeout should be larger than the advisor polling interval. If the timeout is smaller, the manager will ignore reports that logically should be used. By default, advisor reports do not timeout -- the default value is unlimited.
For example, to set the advisor report timeout to 30 seconds for the HTTP advisor for port 80, enter the following command:
ndcontrol advisor timeout http 80 30
For more information on setting the advisor report timeout, see ndcontrol advisor -- control the advisor.
For the Network Dispatcher, you can set the advisor's timeout values at which it detects a server is failed. The failed-server timeout values (connecttimeout and receivetimeout) determine how long an advisor waits before reporting that either a connect or receive has failed.
To obtain the fastest failed-server detection, set the advisor connect and receive timeouts to the smallest value (one second), and set the advisor and manager interval time to the smallest value (one second).
For example, to set the connecttimeout and receivetimeout to 9 seconds for the HTTP advisor on port 80, type the following command:
ndcontrol advisor connecttimeout http 80 9 ndcontrol advisor receivetimeout http 80 9
The default for connect and receive timeout is 3 times the value specified for the advisor interval time.
The custom (customizable) advisor is a small piece of Java code, which you provide as a class file, that gets called by the base code. The base code provides all administrative services, such as starting and stopping an instance of the custom advisor, providing status and reports, and recording history information in a log file. It also reports results to the manager component. Periodically the base code will perform an advisor cycle, where it individually evaluates all servers in its configuration. It starts by opening a connection with a server machine. If the socket opens, the base code will call the "getLoad" method (function) in the custom advisor. The custom advisor then performs whatever steps are necessary to evaluate the health of the server. Typically, it will send a user-defined message to the server and then wait for a response. (Access to the open socket is provided to the custom advisor.) The base code then closes the socket with the server and reports the load information to the Manager.
The base code and custom advisor can operate in either normal or replace mode. Choice of the mode of operation is specified in the custom advisor file as a parameter in the constructor method.
In normal mode, the custom advisor exchanges data with the server, and the base advisor code times the exchange and calculates the load value. The base code then reports this load value to the manager. The custom advisor needs only return a zero (on success) or negative one (on error). To specify normal mode, the replace flag in the constructor is set to false.
In replace mode, the base code does not perform any timing measurements. The custom advisor code performs whatever operations are desired for its unique requirements, and then returns an actual load number. The base code will accept the number and report it to the manager. For best results, normalize your load number between 10 and 1000, with 10 representing a fast server, and 1000 representing a slow server. To specify replace mode, the replace flag in the constructor is set to true.
With this feature, you can write your own advisors that will provide the precise information about servers that you need. A sample custom advisor, ADV_sample.java, is provided with the Network Dispatcher product. After installing Network Dispatcher, you may find the sample code in ...nd/servers/samples/CustomAdvisors install directory.
The default install directories are:
Sample custom advisor files specifically for the WebSphere Application Server advisor are provided in the Network Dispatcher install directory.
The WebSphere Application Server advisor sample files reside in the same samples directory as the ADV_sample.java file.
Your custom advisor file name must be in the form "ADV_myadvisor.java." It must start with the prefix " ADV_" in uppercase. All subsequent characters must be in lowercase letters.
As per Java conventions, the name of the class defined within the file must match the name of the file. If you copy the sample code, be sure to change all instances of "ADV_sample" inside the file to your new class name.
Custom advisors are written in Java language. You must get and install a Java 1.3 compiler for your machine. These files are referenced during compilation:
Your classpath must point to both the custom advisor file and the base classes file during the compile.
For Windows 2000, a compile command might look like this:
javac -classpath <install_dir>\nd\servers\lib\ibmnd.jar ADV_fred.java
where:
The output for the compilation is a class file, for example
ADV_fred.class
Before starting the advisor, copy the class file to the ...nd/servers/lib/CustomAdvisors directory where Network Dispatcher is installed.
For AIX, Linux, and Sun, the syntax is similar.
To run the custom advisor, you must first copy the class file to the proper Network Dispatcher subdirectory:
.../nd/servers/lib/CustomAdvisors/ADV_fred.class
Configure the component, start its manager function, and issue the command to start your custom advisor:
ndcontrol advisor start fred 123
where:
Like all advisors, a custom advisor extends the function of the advisor base, called ADV_Base. It is the advisor base that actually performs most of the advisor's functions, such as reporting loads back to the manager for use in the manager's weight algorithm. The advisor base also performs socket connect and close operations and provides send and receive methods for use by the advisor. The advisor itself is used only for sending and receiving data to and from the port on the server being advised. The TCP methods within the advisor base are timed to calculate the load. A flag within the constructor in the ADV_base overwrites the existing load with the new load returned from the advisor if desired.
These are base class methods:
Network Dispatcher first looks at the list of native advisors that it provides. If it does not find a given advisor there, Network Dispatcher then looks at the customer's list of customized advisors.
/usr/lpp/nd/servers/lib/CustomAdvisors/
/opt/nd/servers/lib/CustomAdvisors/
/opt/nd/servers/lib/CustomAdvisors/
Common install directory path:
C:\Program Files\IBM\edge\nd\servers\lib\CustomAdvisors
Native install directory path:
C:\Program Files\IBM\nd\servers\lib\CustomAdvisors
The program listing for a sample advisor is included in Sample advisor. After installation, this sample advisor can be found in the ...nd/servers/samples/CustomAdvisors directory.
WLM is code that runs on MVS mainframes. It can be queried to ask about the load on the MVS machine.
When MVS Workload Management has been configured on your OS/390 system, Dispatcher can accept capacity information from WLM and use it in the load balancing process. Using the WLM advisor, Dispatcher will periodically open connections through the WLM port on each server in the Dispatcher host table and accept the capacity integers returned. Since these integers represent the amount of capacity that is still available and Dispatcher expects values representing the loads on each machine, the capacity integers are inverted by the advisor and normalized into load values (i.e. a large capacity integer but a small load value both represent a healthier server). The resulting loads are placed into the System column of the manager report.
There are several important differences between the WLM advisor and other Dispatcher advisors:
Like the Metric Server agent, the WLM agent reports on server systems as a whole, rather than on individual protocol-specific server daemons. Metric Server and WLM place their results into the system column of the manager report. As a consequence, running both the WLM advisor and Metric Server at the same time is not supported.
This feature is available for all the Network Dispatcher components.
Metric Server provides server load information to the Network Dispatcher in the form of system-specific metrics, reporting on the health of the servers. The Network Dispatcher manager queries the Metric Server agent residing on each of the servers, assigning weights to the load balancing process using the metrics gathered from the agents. The results are also placed into the manager report.
For a configuration example see Figure 11.
Like the WLM advisor, the Metric Server reports on server systems as a whole, rather than on individual protocol-specific server daemons. Both WLM and Metric Server place their results into the system column of the manager report. As a consequence, running both the WLM advisor and Metric Server at the same time is not supported.
The Metric Server agent must be installed and running on servers that Network Dispatcher is load balancing.
Below are the steps to configure Metric Server for Dispatcher. Similar steps can be used for configuring Metric Server for the other components of Network Dispatcher.
port is the RMI port chosen for all the Metric Server agents to run on. The default RMI port that is set in the metricserver.cmd file is 10004.
systemMetric is the name of the script (residing on the backend server) which should run on each of the servers in the configuration under the specified cluster (or site name). Two scripts are provided for the customer - cpuload and memload. Or, you can create custom system metric scripts. The script contains a command which should return a numeric value in the range of 0-100. This numeric value should represent a load measurement, not an availability value.
Limitation: For Windows 2000, if the name of your System Metric script has an extension other than ".exe", you must specify the full name of the file (for example, "mysystemscript.bat"). This is due to a Java limitation.
Optionally, customers can write their own customized metric script files which define the command that the Metric Server will issue on the server machines. Ensure that any custom scripts are executable and located in the ...nd/ms/script directory. Custom scripts must return a numeric load value in the range of 0-100.
To have Metric Server run on an address other than the local host, you need to edit the metricserver file on the load balanced server machine. After the occurance of "java" in the metricserver file, insert the following:
-Djava.rmi.server.hostname=OTHER_ADDRESS
In addition, before the "if" statements in the metricserver file, add the following line: hostname OTHER_ADDRESS.
For Windows 2000: You will also need to alias the OTHER_ADDRESS on the Microsoft stack. To alias an address on the Microsoft stack, see page ***.
When defining a server in the Network Dispatcher configuration, you are able to distribute the load based on the health of the overall server (using Metric Server agent) and/or the health of any port-specific application (using the advisor function).
With server partitioning, you can further distinguish between particular URLs and their specific applications. For example, one Web server can serve JSP pages, HTML pages, database requests, and so on. Network Dispatcher now provides the ability to partition one cluster and port specific server into several logical servers. This allows you to advise on a particular service on the machine to detect if a servlet engine or a database request is running faster, or not running at all.
Server partitioning allows Network Dispatcher to detect, for example, that the HTML service is serving pages rapidly, but the database connection has gone down. This allows you to distribute load based on more granular service-specific workload, rather than server-wide weighting alone.
Within the Network Dispatcher configuration, you can represent a physical server or a logical server using the cluster:port:server hierarchy. The server can be a unique IP address of the machine (physical server) in either a symbolic name or dotted-decimal format. Or, if you configure the server to represent a partitioned server, then you must provide a resolvable server address for the physical server on the address parameter of the ndcontrol server add command. See ndcontrol server -- configure servers for more information.
Below is an example of partitioning physical servers into logical servers to handle different types of requests.
Cluster: 1.1.1.1 Port: 80 Server: A (IP address 1.1.1.2) html server Server: B (IP address 1.1.1.2) gif server Server: C (IP address 1.1.1.3) html server Server: D (IP address 1.1.1.3) jsp server Server: E (IP address 1.1.1.4) gif server Server: F (IP address 1.1.1.4) jsp server Rule1: \*.htm Server: A Server: C Rule2: \*.jsp Server: D Server: F Rule3: \*.gif Server: B Server: E
In this example, server 1.1.1.2 is partitioned into 2 logical servers -- A (handling html requests) and B (handling gif requests). Server 1.1.1.3 is partitioned into 2 logical servers -- C (handling html requests) and D (handling jsp requests). Server 1.1.1.4 is partitioned into 2 logical servers -- E (handling gif requests) and F (handling jsp requests).
The URL option for the HTTP advisor is available for the Dispatcher and CBR components.
After you have started an HTTP advisor, you can define a unique client HTTP URL string, specific for the service that you want to query on the server. This allows the HTTP advisor to assess the health of the individual services within a server. You can do this by defining logical servers with unique server names that have the same physical IP address. See Server Partitioning: logical servers configured to one physical server (IP address) for more information.
For each defined logical server under the HTTP port you can specify a unique client HTTP URL string, specific for the service that you want to query on the server. The HTTP advisor uses the advisorrequest string to query the health of the servers. The default value is HEAD / HTTP/1.0. The advisorresponse string is the advisor response that the HTTP advisor scans for in the HTTP response. The HTTP advisor uses the advisorresponse string to compare to the real response that is received from the server. The default value is null.
Important: If a blank is contained within the HTTP URL string:
server set cluster:port:server advisorrequest "head / http/2.0" server set cluster:port:server advisorresponse "HTTP 200 OK"
ndcontrol server set cluster:port:server advisorrequest "\"head / http/2.0\"" ndcontrol server set cluster:port:server advisorresponse "\"HTTP 200 OK\""
Network Dispatcher can reside on the same machine as a server for which it is load balancing requests. This is commonly referred to as collocating a server. Collocation applies to the Dispatcher, Site Selector, Mailbox Locator, and Cisco Consultant components. Collocation is also supported for CBR, but only when using bind-specific Web servers and bind-specific Caching Proxy.
Red Hat Linux v7.1 (Linux kernel version 2.4.2-2) or SuSE Linux v7.1 (Linux kernel version 2.4.0-4GB): In order to configure both collocation and high availability at the same time, when running the Dispatcher component using the mac forwarding method, you must install a Linux kernel patch. For more information on installing the patch, see Installing the Linux kernel patch (to suppress arp responses on the loopback interface). However, when following these instructions, skip the step to alias the loopback adapter. You should add the ifconfig instruction to alias the loopback adapter in the goStandby high-availability script file that gets executed when a Dispatcher goes into standby state.
Solaris: There is a limitation that you cannot configure WAND advisors when the entry-point Dispatcher is collocated. See Using remote advisors with wide area support.
In earlier releases, it was necessary to specify the collocated server address to be the same as the nonforwarding address (NFA) in the configuration. That restriction has been lifted.
To configure a server to be collocated, the ndcontrol server command provides an option called collocated which can be set to yes or no. The default is no. The address of the server must be a valid IP address of a network interface card on the machine.
You can configure a collocated server in one of the following ways:
See ndcontrol server -- configure servers, for more information on ndcontrol server command syntax.
CBR supports collocation on all platforms with no additional configurations required. However, the Web servers and Caching Proxy that you use must be bind-specific.
Mailbox Locator supports collocation on all platforms. However, the server must be bound to a different address than Network Dispatcher for this to work. In order to collocate a POP3 or IMAP server on the same machine it must be bound to an IP address which is different from the cluster address. This can be achieved through the use of the loopback address.
Site Selector supports collocation on all platforms with no additional configurations required.
Cisco Consultant supports collocation on all platforms with no additional configurations required.
This feature is only available for the Dispatcher component.
If you are not using the Dispatcher's wide area support and not using Dispatcher's nat forwarding method, a Dispatcher configuration requires that the Dispatcher machine and its servers all be attached to the same LAN segment (see Figure 22). A client's packet comes into the ND machine and is sent to the server, and then from the server directly back to the client.
Figure 22. Example of a configuration consisting of a single LAN segment
The Wide Area Dispatcher enhancement adds support for offsite servers, known as remote servers (see Figure 23). If GRE is not supported at the remote site and you are not using Dispatcher's nat forwarding method, then the remote site must consists of a remote Dispatcher machine (Dispatcher 2) and its locally attached servers (ServerG, ServerH, and ServerI). All the Dispatcher machines must be on the same operating system. A client's packet can now go from the Internet to a Dispatcher machine, from there to a geographically remote Dispatcher machine to one of its locally attached servers.
Figure 23. Example of configuration using local and remote servers
This allows one cluster address to support all worldwide client requests while distributing the load to servers around the world.
The Dispatcher machine initially receiving the packet can still have local servers attached to it and it can distribute the load between its local servers and the remote servers.
Wide area commands are not complex. To configure wide area support :
ndcontrol server add cluster:port:server router address
For more information on the router keyword, see ndcontrol server -- configure servers.
On entry-point Dispatchers, advisors will work correctly without any special configuration for most platforms.
Linux: There is a limitation on using remote advisors with wide area support configurations. Protocol-specific advisors, such as the HTTP advisor, that are running on the entry-point Dispatcher machine will not correctly assess the status of server machines at the remote site. To alleviate this problem, do one of the following:
Either of these options will provide the advisor running on the entry point Dispatcher machine with an assessment of the status of the remote Dispatcher machine.
Solaris: On entry-point Network Dispatchers, you must use the arp configuration method (instead of the ifconfig or cluster configuration methods). For example:
arp -s <my_cluster_address> <my_mac_address> pub
On remote Dispatchers, you will need to perform the following configuration steps for each remote cluster address. For a high-availability configuration at the remote Network Dispatcher location, you must perform these steps on both machines.
AIX
ifconfig lo0 alias 9.67.34.123 netmask 255.255.255.255
Linux
ifconfig lo:1 9.67.34.123 netmask 255.255.255.255 up
Solaris
Windows 2000
ndconfig en0 alias 9.55.30.45 netmask 255.255.240.0
arp -a
arp -d 9.67.34.123
and look for the address of your machine.
route add 9.67.34.123 mask 255.255.255.255 9.55.30.45
Figure 24. Wide area example configuration with remote Network Dispatchers
This example applies to the configuration illustrated in Figure 24.
Here is how to configure the Dispatcher machines to support cluster address xebec on port 80. ND1 is defined as the "entry point,". An ethernet connection is assumed. Note that ND1 has five servers defined: three local (ServerA, ServerB, ServerC) and two remote (ND2 and ND3). Remotes ND2 and ND3 each have three local servers defined.
At the console of the first Dispatcher (ND1), do the following:
ndcontrol executor start
ndcontrol executor set nfa ND1
ndcontrol cluster add xebec
ndcontrol port add xebec:80
ndcontrol cluster configure ND1 and also configure xebec as clusteraddr.
ndcontrol cluster configure xebec
At the console of the second Dispatcher (ND2):
ndcontrol executor start
ndcontrol executor set nfa ND2
ndcontrol cluster add xebec
ndcontrol port add xebec:80
ndcontrol cluster configure ND2
At the console of the third Dispatcher (ND3):
ndcontrol executor start
ndcontrol executor set nfa ND3
ndcontrol cluster add xebec
ndcontrol port add xebec:80
ndcontrol cluster configure ND3
Generic Routing Encapsulation (GRE) is an internet protocol specified in RFC 1701 and RFC 1702. Using GRE, the Network Dispatcher can encapsulate client IP packets inside IP/GRE packets and forward them to server platforms such as OS/390 that support GRE. GRE support allows the Dispatcher component to load balance packets to multiple server addresses associated with one MAC address.
Network Dispatcher implements GRE as part of its WAND (wide area Network Dispatcher) feature. This allows Network Dispatcher to provide wide area load balancing directly to any server systems that can unwrap the GRE packets. Network Dispatcher does not need to be installed at the remote site if the remote servers support the encapsulated GRE packets. Network Dispatcher encapsulates WAND packets with the GRE key field set to decimal value 3735928559.
Figure 25. Wide area example configuration with server platform that supports GRE
For this example (Figure 25), to add remote ServerD, which supports GRE, define it within your Network Dispatcher configuration as if you are defining a WAND server in the cluster:port:server hierarchy:
ndcontrol server add cluster:port:ServerD router Router1
The self advisor is available on the Dispatcher component.
For Network Dispatcher in a two-tiered WAND (wide area Network Dispatcher) configuration, Dispatcher provides a self advisor that collects load status information on backend servers.
Figure 26. Example of a two-tiered WAND configuration using the self advisor
In this example, the self advisor along with Metric Server reside on the two Dispatcher machines that are being load balanced by the top tier Network Dispatcher. The self advisor specifically measures the connections per second rate on backend servers of the Dispatcher at the executor level.
The self advisor writes the results to the ndloadstat file. Network Dispatcher also provides an external metric called ndload. The Metric Server agent on each Dispatcher machine runs its configuration that calls the external metric ndload. The ndload script extracts a string from the ndloadstat file and returns it to the Metric Server agent. Subsequently, each of the Metric Server agents (from each of the Dispatchers) returns the load status value to the top-tiered Network Dispatcher for use in determining which Dispatcher to forward client requests.
The ndload executable resides in the .../nd/ms/script directory for Network Dispatcher.
The high availability feature is only available for the Dispatcher component.
To improve Dispatcher availability, the Dispatcher high availability function uses the following mechanisms:
If possible, it is recommended that at least one of the heartbeat pairs be across a separate subnet than the regular cluster traffic. Keeping the heartbeat traffic distinct will help prevent false takeovers during very heavy network loads and also improve complete recovery times after a failover.
Complete syntax for ndcontrol highavailability is in ndcontrol highavailability -- control high availability.
For a more complete discussion of many of the tasks below, see Setting up the Dispatcher machine.
Windows 2000 only: In addition, configure each nonforwarding address using the ndconfig command. For example:
ndconfig en0 nfa_addr netmask netmask
ndcontrol cluster set clusterA primaryhost NFAdispatcher1 ndcontrol cluster set clusterB primaryhost NFAdispatcher2
ndcontrol cluster set clusterB primaryhost NFAdispatcher2 ndcontrol cluster set clusterA primaryhost NFAdispatcher1
ndcontrol highavailability heartbeat add sourceaddress destinationaddress
Primary - highavailability heartbeat add 9.67.111.3 9.67.186.8 Backup - highavailability heartbeat add 9.67.186.8 9.67.111.3At least one heartbeat pair must have the NFAs of the pair as the source and destination address.
If possible, it is recommended that at least one of the heartbeat pairs be across a separate subnet than the regular cluster traffic. Keeping the heartbeat traffic distinct will help prevent false takeovers during very heavy network loads and also improve complete recovery times after a failover.
ndcontrol highavailability reach add 9.67.125.18Reach targets are recommended but not required. See Failure detection capability using heartbeat and reach target, for more information.
ndcontrol highavailability backup add primary [auto | manual] port
ndcontrol highavailability backup add backup [auto | manual] port
ndcontrol highavailability backup add both [auto | manual] port
ndcontrol highavailability statusThe machines should each have the correct role (backup, primary, or both), states, and substates. The primary should be active and synchronized; the backup should be in standby mode and should be synchronized within a short time. The strategies must be the same.
Notes:
Besides the basic criteria of failure detection (the loss of connectivity between active and standby Dispatchers, detected through the heartbeat messages), there is another failure detection mechanism named reachability criteria. When you configure the Dispatcher you can provide a list of hosts that each of the Dispatchers should be able to reach in order to work correctly.
You should choose at least one host for each subnet your Dispatcher machine uses. The hosts could be routers, IP servers or other types of hosts. Host reachability is obtained by the reach advisor, which pings the host. Switchover takes place either if the heartbeat messages cannot go through, or if the reachability criteria are met better by the standby Dispatcher than by the primary Dispatcher. To make the decision based on all available information, the active Dispatcher regularly sends the standby Dispatcher its reachability capabilities. The standby Dispatcher then compares those capabilities with its own and decides whether to switch.
Two Dispatcher machines are configured: the primary machine, and a second machine called the backup. At startup, the primary machine sends all the connection data to the backup machine until that machine is synchronized. The primary machine becomes active, that is, it begins load balancing. The backup machine, meanwhile, monitors the status of the primary machine, and is said to be in standby state.
If the backup machine at any point detects that the primary machine has failed, it performs a takeover of the primary machine's load balancing functions and becomes the active machine. After the primary machine has once again become operational, the machines respond according to how the recovery strategy has been configured by the user. There are two kinds of strategy:
The strategy parameter must be set the same for both machines.
The manual recovery strategy allows you to force the routing of packets to a particular machine, using the takeover command. Manual recovery is useful when maintenance is being performed on the other machine. The automatic recovery strategy is designed for normal unattended operation.
For a mutual high availability configuration, there is no per cluster failure. If any problem occurs with one machine, even if it affects just one cluster, then the other machine will take over for both clusters.
For Dispatcher to route packets, each cluster address must be aliased to a network interface device.
Since the Dispatcher machines will change states when a failure is detected, the commands above must be issued automatically. Dispatcher will execute user-created scripts to do that. Sample scripts can be found in the ...nd/servers/samples directory and must be moved to the ...nd/servers/bin directory in order to run.
The following sample scripts may be used:
call netsh interface ip add address "Local Area Connection" addr=9.37.51.28 mask=255.255.240.0
In the goStandby and GoInOp, the alias will need to be removed. For example:
call netsh interface ip delete address "Local Area Connection" addr=9.37.51.28
If there are multiple NIC's on the machine, then first check which interface you should use by issuing the following command on the command prompt: netsh interface ip show address. This command will return a list of currently configured interfaces and will number the "Local Area Connection" (for example, "Local Area Connection 2") so you can determine which one you should use.
You can use rules-based load balancing to fine tune when and why packets are sent to which servers. Network Dispatcher reviews any rules you add from first priority to last priority, stopping on the first rule that it finds to be true, then load balancing the content between any servers associated with the rule. It already balances the load based on destination and port, but using rules expands your ability to distribute connections.
In most cases when configuring rules, you should configure a default always true rule in order to catch any request that falls through the other higher priority rules. This can be a "Sorry, the site is currently down, please try again later" response when all other servers fail for the client request.
You should use rules-based load balancing with Dispatcher and Site Selector when you want to use a subset of your servers for some reason. You must always use rules for the CBR component.
You can choose from the following types of rules:
We recommend that you make a plan of the logic that you want the rules to follow before you start adding rules to your configuration.
All rules have a name, type, priority, and may have a begin range and end range, along with a set of servers. In addition, the content type rule for the CBR component has a matching regular expression pattern associated with it. ( For examples and scenarios on how to use the content rule and valid pattern syntax for the content rule, see Appendix C, Content rule (pattern) syntax.)
Rules are evaluated in priority order. In other words, a rule with a priority of 1 (lower number) will be evaluated before a rule with a priority of 2 (higher number). The first rule that is satisfied will be used. Once a rule has been satisfied, no further rules are evaluated.
For a rule to be satisfied, it must meet two conditions:
If a rule has no servers associated with it, the rule only needs to meet condition one to be satisfied. In this case, Dispatcher will drop the connection request, Site Selector will return the name server request with an error, and CBR will cause Caching Proxy to return an error page.
If no rules are satisfied, Dispatcher will select a server from the full set of servers available on the port, Site Selector will select a server from the full set of servers available on the site name, and CBR will cause Caching Proxy to return an error page.
This rule type is available in the Dispatcher, CBR, or Site Selector component.
You may want to use rules based on the client IP address if you want to screen the customers and allocate resources based on where they are coming from.
For example, you have noticed that your network is getting a lot of unpaid and therefore unwanted traffic from clients coming from a specific set of IP addresses. You create a rule using the ndcontrol rule command, for example:
ndcontrol rule add 9.67.131.153:80:ni type ip beginrange 9.0.0.0 endrange 9.255.255.255
This "ni" rule would screen out any connection from IBM clients. You would then add to the rule the servers which you wanted accessible to IBMers, or if you do not add any servers to the rule, requests coming from 9.x.x.x addresses would not be served by any of your servers.
This rule type is available in the Dispatcher, CBR, or Site Selector component.
You may want to use rules based on the time of day for capacity planning reasons. For example, if your Web site gets hit most during the same group of hours every day, you might want to dedicate five servers to HTTP during full-time, then adding another five during the peak time period.
Another reason you might use a rule based on the time of day is when you want to take some of the servers down for maintenance every night at midnight, so you would set up a rule that excludes those servers during the necessary maintenance period.
This rule type is available in the Dispatcher and CBR component.
You may want to use rules based on connections per second on a port if you need to share some of your servers with other applications. For example, you can set two rules:
Or you might be using Telnet and want to reserve two of your five servers for Telnet, except when the connections per second increase above a certain level. This way, Dispatcher would balance the load across all five servers at peak times.
This rule type is available in the Dispatcher or CBR component.
You may want to use rules based on active connections total on a port if your servers get overloaded and start throwing packets away. Certain Web servers will continue to accept connections even though they do not have enough threads to respond to the request. As a result, the client requests time out and the customer coming to your Web site is not served. You can use rules based on active connections to balance capacity within a pool of servers.
For example, you know from experience that your servers will stop serving after they have accepted 250 connections. You can create a rule using the ndcontrol rule command or the cbrcontrol rule command, for example:
ndcontrol rule add 130.40.52.153:80:pool2 type active beginrange 250 endrange 500 or cbrcontrol rule add 130.40.52.153:80:pool2 type active beginrange 250 endrange 500
You would then add to the rule your current servers plus some additional servers, which will otherwise be used for other processing.
This rule type is only available in the Dispatcher component.
You may want to use rules based on the client port if your clients are using some kind of software that asks for a specific port from TCP/IP when making requests.
For example, you could create a rule that says that any request with a client port of 10002 will get to use a set of special fast servers because you know that any client request with that port is coming from an elite group of customers.
This rule type is only available in the Dispatcher component.
You may want to use rules based on the content of the "type of service" (TOS) field in the IP header. For example, if a client request comes in with one TOS value that indicates normal service, it can be routed to one set of servers. If a different client request comes in with a different TOS value that indicates a higher priority of service, it can be routed to a different set of servers.
The TOS rule allows you to fully configure each bit in the TOS byte using the ndcontrol rule command. For significant bits that you want matched in the TOS byte, use 0 or 1. Otherwise, the value x is used. The following is an example for adding a TOS rule:
ndcontrol rule add 9.67.131.153:80:tsr type service tos 0xx1010x
Capacity utilization and bandwidth rules are only available in the Dispatcher component.
Using the capacity utilization feature, Dispatcher measures the amount of data delivered by each of its servers. Dispatcher tracks capacity at the server, rule, port, cluster, and executor levels. For each of these levels, there is a new byte counter value: kilobytes transferred per second. The rate value (kilobytes transferred per second) is calculated over a 60 second interval. You can view these capacity values from the GUI or from the output of a command line report.
Dispatcher allows you to allocate a specified bandwidth to sets of servers within your configuration using the reserved bandwidth rule. When traffic exceeds the reserved bandwidth threshold, you can do either of the following:
By using the shared bandwidth rule in conjunction with the reserved bandwidth rule, as described above, you can provide preferred clients with increased server access and therefore optimize performance for their transactions. For example, using the shared bandwidth rule to recruit unused bandwidth, you can allow online trading customers executing trades on server clusters to receive greater access than customers using other server clusters for investment research.
Note the following to determine whether bandwidth rules can help you manage the volume of response traffic that flows from servers to clients:
This rule type is only available in the Dispatcher component.
The reserved bandwidth rule allows you to load balance based on the number of kilobytes per second being delivered by a set of servers. By setting a threshold (allocating a specified bandwidth range) for each set of servers throughout the configuration, you can control and guarantee the amount of bandwidth being used by each cluster-port combination. The following is an example for adding a reservedbandwidth rule:
ndcontrol rule add 9.67.131.153:80:rbw type reservedbandwidth beginrange 0 endrange 300
The begin range and endrange are specified in kilobytes per second.
This rule type is only available in the Dispatcher component.
If the amount of data transferred exceeds the limit for the reserved bandwidth rule, the shared bandwidth rule provides you the ability to recruit unused bandwidth available at the site. You can configure this rule to share bandwidth at either the cluster or the executor level. Sharing bandwidth at the cluster level allows a port (or ports) to share a maximum amount of bandwidth across several ports (applications/ protocols) within the same cluster. Sharing bandwidth at the executor level allows a cluster (or clusters) within the entire Dispatcher configuration to share a maximum amount of bandwidth.
Prior to configuring the shared bandwidth rule, you must specify the maximum amount of bandwidth (kilobytes per second) that can be shared at the executor or cluster level using ndcontrol executor or ndcontrol cluster command with the sharedbandwidth option. The following are examples of the command syntax:
ndcontrol executor set sharedbandwidth size ndcontrol cluster [add | set] 9.12.32.9 sharedbandwidth size
The size for sharedbandwidth is an integer value (kilobytes per second). The default is zero. If the value is zero, then bandwidth cannot be shared. You should specify a maximum sharedbandwidth value that does not exceed the total bandwidth (total server capacity) available.
The following are examples of adding or setting a sharedbandwidth rule:
ndcontrol rule add 9.20.30.4:80:shbw type sharedbandwidth sharelevel value ndcontrol rule set 9.20.34.11:80:shrule sharelevel value
The value for sharelevel is either executor or cluster. Sharelevel is a required parameter on the sharebandwidth rule.
This rule type is only available in the Site Selector component.
For the metric all rule, you choose a system metric (cpuload, memload, or your own customized system metric script), and Site Selector compares the system metric value (returned by the Metric Server agent residing in each load-balanced server) with the begin and end range that you specify in the rule. The current system metric value for all the servers in the server set must be within the range for the rule to fire.
The following is an example of adding a metric all rule to your configuration:
sscontrol rule add dnsload.com:allrule1 type metricall metricname cpuload beginrange 0 endrange 100
This rule type is only available in the Site Selector component.
For the metric average rule, you choose a system metric (cpuload, memload, or your own customized system metric script), and Site Selector compares the system metric value (returned by the Metric Server agent residing in each load-balanced server) with the begin and end range that you specify in the rule. The average of the current system metric values for all the servers in the server set must be within the range for the rule to fire.
The following is an example of adding a metric average rule to your configuration:
sscontrol rule add dnsload.com:avgrule1 type metricavg metricname cpuload beginrange 0 endrange 100
This rule type is available in the Dispatcher, CBR, or Site Selector component.
A rule may be created that is "always true." Such a rule will always be selected, unless all the servers associated with it are down. For this reason, it should ordinarily be at a lower priority than other rules.
You can even have multiple "always true" rules, each with a set of servers associated with it. The first true rule with an available server is chosen. For example, assume you have six servers. You want two of them to handle your traffic under all circumstances, unless they are both down. If the first two servers are down, you want a second set of servers to handle the traffic. If all four of these servers are down, then you will use the final two servers to handle the traffic. You could set up three "always true" rules. Then the first set of servers will always be chosen as long as at least one is up. If they are both down, one from the second set will be chosen, and so forth.
As another example, you may want an "always true" rule to ensure that if incoming clients do not match any of the rules you have set, they will not be served. You would create a rule using the ndcontrol rule command like:
ndcontrol rule add 130.40.52.153:80:jamais type true priority 100
Then you would not add any servers to the rule, causing the clients packets to be dropped with no response.
You can define more than one "always true" rule, and thereafter adjust which one gets executed by changing their priority levels.
This rule type is available in the Dispatcher or CBR component.
You will want to use content type rules to send requests to sets of servers specifically set up to handle some subset of your site's traffic. For example, you may want to use one set of servers to handle all cgi-bin requests, another set to handle all streaming audio requests, and a third set to handle all other requests. You would add one rule with a pattern that matches the path to your cgi-bin directory, another that matches the file type of your streaming audio files, and a third always true rule to handle the rest of the traffic. You would then add the appropriate servers to each of the rules.
Important: For examples and scenarios on how to use the content rule and valid pattern syntax for the content rule, see Appendix C, Content rule (pattern) syntax.
You can add rules using the ndcontrol rule add command, by editing the sample configuration file, or with the graphical user interface (GUI). You can add one or more rules to every port you have defined.
It is a two-step process: add the rule, then define which servers to serve to if the rule is true. For example, our system administrator wanted to track how much use the proxy servers were getting from each division on site. She knows which IP addresses are given to each division. She would create the first set of rules based on client IP address to separate each division's load:
ndcontrol rule add 130.40.52.153:80:div1 type ip b 9.1.0.0 e 9.1.255.255 ndcontrol rule add 130.40.52.153:80:div2 type ip b 9.2.0.0 e 9.2.255.255 ndcontrol rule add 130.40.52.153:80:div3 type ip b 9.3.0.0 e 9.3.255.255
Next, she would add a different server to each rule, then measure the load on each of the servers in order to bill the division properly to the services they are using. For example:
ndcontrol rule useserver 130.40.52.153:80:div1 207.72.33.45 ndcontrol rule useserver 130.40.52.153:80:div2 207.72.33.63 ndcontrol rule useserver 130.40.52.153:80:div3 207.72.33.47
The server evaluation option is only available in the Dispatcher component.
On the ndcontrol rule command there is a server evaluation option for rules. Use the evaluate option to choose to evaluate the rule's condition across all the servers on the port or to evaluate the rule's condition across just the servers within the rule. (In earlier versions of Network Dispatcher, you could only measure each rule's condition across all servers on the port.)
The following are examples of adding or setting the evaluate option on a reserved bandwidth rule:
ndcontrol rule add 9.22.21.3:80:rbweval type reservedbandwidth evaluate level ndcontrol rule set 9.22.21.3:80:rbweval evaluate level
The evaluate level can be set to either port or rule. The default is port.
The option to measure the rule's condition across the servers within the rule allows you to configure two rules with the following characteristics:
The result is that when traffic exceeds the threshold of the servers within the first rule, traffic will be sent to the "site busy" server within the second rule. When traffic falls below the threshold of the servers within the first rule, new traffic continues once again to the servers in the first rule.
Using the two rules described in the previous example, if you set the evaluate option to port for the first rule (evaluate rule's condition across all the servers on the port), when traffic exceeds the threshold of that rule, traffic is sent to the "site busy" server associated to the second rule.
The first rule measures all server traffic (including the "site busy" server) on the port to determine whether the traffic exceeds the threshold. As congestion decreases for the servers associated to the first rule, an unintentional result may occur where traffic continues to the "site busy" server because traffic on the port still exceeds the threshold of the first rule.
In general, the load-balancing functions of the Dispatcher work independently of the content of the sites on which the product is used. There is one area, however, where site content can be important, and where decisions made regarding content can have a significant impact upon the Dispatcher's efficiency. This is in the area of link addressing.
If your pages specify links that point to individual servers for your site, you are in effect forcing a client to go to a specific machine, thus bypassing any load balancing function that might otherwise be in effect. For this reason, it is recommended that you always use the address of Dispatcher in any links contained in your pages. Note that the kind of addressing used may not always be apparent, if your site uses automated programming that dynamically creates HTML. To maximize your load-balancing, you should be aware of any explicit addressing and avoid it where possible.
You can set up Dispatcher and the TCP server machines using a private network. This configuration can reduce the contention on the public or external network that can affect performance.
For AIX, this configuration can also take advantage of the fast speeds of the SP High Performance Switch if you are running Dispatcher and the TCP server machines on nodes in an SP Frame.
To create a private network, each machine must have at least two LAN cards, with one of the cards connected to the private network. You must also configure the second LAN card on a different subnet. The Dispatcher machine will then send the client requests to the TCP server machines through the private network.
Windows 2000: Execute the following command:
ndconfig en1 10.0.0.x netmask 255.255.255.0
Where en1 is the name of the second interface card in the Dispatcher machine, 10.0.0.x is the network address of the second interface card, and 255.255.255.0 is the netmask of the private network.
The servers added using the ndcontrol server add command must be added using the private network addresses; for example, referring to the Apple server example in Figure 27, the command should be coded as:
ndcontrol server add cluster_address:80:10.0.0.1
not
ndcontrol server add cluster_address:80:9.67.131.18
If you are using Site Selector to provide load information to Dispatcher, you must configure Site Selector to report loads on the private addresses.
Figure 27. Example of a private network using Dispatcher
Using a private network configuration only applies to the Dispatcher component.
The "wildcard" refers to the cluster's ability to match multiple IP addresses (i.e. acts as a wildcard). Cluster address 0.0.0.0 is used to specify a wildcard cluster.
If you have many cluster addresses to load-balance, and the port/server configurations are identical for all your clusters, you can combine all the clusters into one star configuration.
You must still explicitly configure each cluster address on one of the network adapters of your Dispatcher workstation. You should not add any of the cluster addresses to the Dispatcher configuration using the ndcontrol cluster add command however.
Add only the wildcard cluster (address 0.0.0.0), and configure the ports and servers as required for load balancing. Any traffic to any of the adapter configured addresses will be load balanced using the wildcard cluster configuration.
An advantage of this approach is that traffic to all the cluster addresses is taken into account when determining the best server to go to. If one cluster is getting a lot of traffic, and it has created many active connections on one of the servers, traffic to other cluster addresses will be load balanced using this information.
You can combine the wildcard cluster with actual clusters if you have some cluster addresses with unique port/server configurations, and some with common configurations. The unique configurations must each be assigned to an actual cluster address. All common configurations can be assigned to the wildcard cluster.
Using wildcard cluster to combine server configurations only applies to the Dispatcher component.
Using wildcard cluster to load balance firewalls only applies to the Dispatcher component. Cluster address 0.0.0.0 is used to specify a wildcard cluster.
The wildcard cluster can be used to load balance traffic to addresses that are not explicitly configured on any network adapter of the Dispatcher workstation. In order for this to work, the Dispatcher must at least be able to see all the traffic it is to load balance. The dispatcher workstation will not see traffic to addresses that have not been explicitly configured on one of its network adapters unless it is set up as the default route for some set of traffic.
Once Dispatcher has been configured as a default route, any TCP or UDP traffic through the Dispatcher machine will be load balanced using the wildcard cluster configuration.
One application of this is to load balance firewalls. Since firewalls can process packets for any destination address and any destination port, you need to be able to load balance traffic independent of the destination address and port.
Firewalls are used to handle traffic from non-secure clients to secure servers, and the responses from the secure servers, as well as traffic from clients on the secure side to servers on the non-secure side, and the responses.
You must set up two Dispatcher machines, one to load balance non-secure traffic to the non-secure firewall addresses and one to load balance secure traffic to the secure firewall addresses. Since both of these Dispatchers must use the wildcard cluster and wildcard port with different sets of server addresses, the two Dispatchers must be on two separate workstations.
Using wildcard cluster with Caching Proxy for transparent proxy only applies to the Dispatcher component. Cluster address 0.0.0.0 is used to specify a wildcard cluster.
The wildcard cluster function also allows Dispatcher to be used to enable a transparent proxy function for a Caching Proxy server residing on the same box as Dispatcher. This is an AIX feature only, as there must be communication from the dispatcher component to the TCP component of the operating system.
To enable this feature, you must start Caching Proxy listening for client requests on port 80. You then configure a wildcard cluster. In the wildcard cluster, you configure port 80. In port 80, you configure the NFA of the Dispatcher machine as the only server. Now any client traffic to any address on port 80 will be delivered to the Caching Proxy server running on the Dispatcher workstation. The client request will then be proxied as usual, and the response will be sent back from Caching Proxy to the client. In this mode, the Dispatcher component is not performing any load balancing.
The wildcard port can be used to handle traffic that is not for any explicitly configured port. One use of this is for load balancing firewalls. A second use is to ensure that traffic to an unconfigured port is handled appropriately. By defining a wildcard port with no servers, you will guarantee that any request to a port that has not been configured will be discarded rather than delivered back to the operating system. Port number 0 (zero) is used to specify a wildcard port, for example:
ndcontrol port add cluster:0
You enable the affinity feature when you configure a cluster's port to be sticky. Configuring a cluster's port to be sticky allows subsequent client requests to be directed to the same server. This is done by setting "port stickytime" to some number of seconds. The feature is disabled by setting stickytime to zero.
Interaction with cross port affinity: If you are enabling cross port affinity, stickytime values of the shared ports must be the same (nonzero) value. See Cross port affinity, for more information.
With the feature disabled, whenever a new TCP connection is received from a client, Dispatcher picks the right server at that moment in time and forwards the packets to it. If a subsequent connection comes in from the same client, Dispatcher treats it as an unrelated new connection, and again picks the right server at that moment in time.
With the feature enabled, if a subsequent request is received from the same client, the request is directed to the same server.
Over time, the client will finish sending transactions, and the affinity record will go away. Hence the meaning of the sticky "time. " Each affinity record lives for the "stickytime" in seconds. When subsequent connections are received within the stickytime, the affinity record is still valid and the request will go to the same server. If a subsequent connection is not received within stickytime, the record is purged; a connection that is received after that time will have a new server selected for it.
The Server Directed Affinity API only applies to the Dispatcher component.
The SDA feature provides an API that allows an external agent to influence the Dispatcher affinity behavior.
SDA Features
Your application may have indicated that their server systems have the knowledge to direct client requests to particular server machines better than Dispatcher can. Rather than having a client be "directed" to the same server as selected by Dispatcher's load balancing selection, you may want the client to be "directed" to the server of your choosing. The SDA feature provides this API. You can now write your own software to implement an SDA agent, which communicates with a listener in Dispatcher. It can then manipulate the Dispatcher affinity tables to:
Records inserted in an affinity table by an SDA agent remain in the table indefinitely. They do not timeout. They are removed only when the SDA agent removes them or if a Dispatcher advisor detects that the server is dead.
Dispatcher's SDA components
Dispatcher implements a new socket listener to accept and handle requests from an SDA agent. When an SDA agent opens a connection with Dispatcher, the listener will accept it and leave the connection open. Multiple requests and responses can flow over this persistent connection. The socket will close when the SDA agent closes it or if Dispatcher detects an unrecoverable error. Inside Dispatcher, the listener takes each request from the SDA agent, communicates with the appropriate affinity table in the Dispatcher executor kernel, and prepares a response for the SDA agent.
For more information, refer to the files appearing in the Network Dispatcher's install directory:
Cross port affinity only applies to the Dispatcher component.
Cross port affinity is the sticky feature that has been expanded to cover multiple ports. For example, if a client request is first received on one port and the next request is received on another port, cross port affinity allows the dispatcher to send the client request to the same server. In order to use this feature, the ports must:
More than one port can link to the same crossport. When subsequent connections come in from the same client on the same port or a shared port, the same server will be accessed. The following is an example of configuring multiple ports with a cross port affinity to port 10:
ndcontrol port set cluster:20 crossport 10 ndcontrol port set cluster:30 crossport 10 ndcontrol port set cluster:40 crossport 10
After cross port affinity has been established, you have the flexibility to modify the stickytime value for the port. However, it is recommended that you change the stickytime values for all shared ports to the same value, otherwise unexpected results may occur.
To remove the cross port affinity, set the crossport value back to its own port number. See ndcontrol port -- configure ports, for detailed information on command syntax for the crossport option.
Affinity address mask only applies to the Dispatcher component.
Affinity address mask is a sticky feature enhancement to group clients based upon common subnet addresses. Specifying stickymask on the ndcontrol port command allows you to mask the common high-order bits of the 32-bit IP address. If this feature is enabled, when a client request first makes a connection to the port, all subsequent requests from clients with the same subnet address (represented by that part of the address which is being masked) will be directed to the same server.
For example, if you want all incoming client requests with the same network Class A address to be directed to the same server, you set the stickymask value to 8 (bits) for the port. To group client requests with the same network Class B address, set the stickymask value to 16 (bits). To group client requests with the same network Class C address, set the stickymask value to 24 (bits).
For best results, set the stickymask value when first starting the Network Dispatcher. If you change the stickymask value dynamically, results will be unpredictable.
Interaction with cross port affinity: If you are enabling cross port affinity, stickymask values of the shared ports must be the same. See Cross port affinity, for more information.
To enable affinity address mask, issue an ndcontrol port command similar to the following:
ndcontrol port set cluster:port stickymask 8
Possible stickymask values are 8, 16, 24 and 32. A value of 8 specifies the first 8 high-order bits of the IP address (network Class A address) will be masked. A value of 16 specifies the first 16 high-order bits of the IP address (network Class B address) will be masked. A value of 24 specifies the first 24 high-order bits of the IP address (network Class C address) will be masked. If you specify a value of 32, you are masking the entire IP address which effectively disables the affinity address mask feature. The default value of stickymask is 32.
See ndcontrol port -- configure ports, for detailed information on command syntax for stickymask (affinity address mask feature).
With rule affinity override, you can override the stickiness of a port for a specific server. For example, you are using a rule to limit the amount of connections to each application server, and you have an overflow server with an always true rule that says "please try again later" for that application. The port has a stickytime value of 25 minutes, so you don't want the client to be sticky to that server. With rule affinity override, you can change the overflow server to override the affinity normally associated with that port. The next time the client requests the cluster, it is load balanced to the best available application server, not the overflow server.
See ndcontrol server -- configure servers, for detailed information on command syntax for the rule affinity override, using the server sticky option.
Quiesce handling for sticky connections applies to the Dispatcher and CBR components.
To remove a server from the Network Dispatcher configuration for any reason (updates, upgrades, service, and so forth), you can use the ndcontrol manager quiesce command. The quiesce subcommand allows existing connections to complete (without being severed) and forwards only subsequent new connections from the client to the quiesced server if the connection is designated as sticky and stickytime has not expired. The quiesce subcommand disallows any other new connections to the server.
Only use quiesce "now" if you have stickytime set, and you want new connections sent to another server (instead of the quiesced server) before stickytime expires. The following is an example of using the now option to quiesce server 9.40.25.67:
ndcontrol manager quiesce 9.40.25.67 now
The now option determines how sticky connections will be handled as follows:
This is the more graceful, less abrupt, way to quiesce servers. For instance, you can gracefully quiesce a server and then wait for the time where there is the least amount of traffic (perhaps early morning) to completely remove the server from the configuration.
You can specify the following types of affinity on the ndcontrol rule command:
The default for the affinity option is "none." The stickytime option on the port command must be zero (not enabled) in order to set the affinity option on the rule command to active cookie, passive cookie, or URI. When affinity is set on the rule, you cannot enable stickytime on the port.
The active cookie affinity only applies to the CBR component. The passive cookie and URI affinity apply to the CBR component and to Dispatcher component's cbr forwarding method.
The active cookie affinity feature applies only to the CBR component. It provides a way to make clients "sticky" to a particular server. This function is enabled by setting the stickytime of a rule to a positive number, and setting the affinity to "activecookie." This can be done when the rule is added, or using the rule set command. See ndcontrol rule -- configure rules, for detailed information on command syntax.
Once a rule has been enabled for active cookie affinity, new client requests will be load-balanced using standard CBR algorithms, while succeeding requests from the same client will be sent to the initially chosen server. The chosen server is stored as a cookie in the response to the client. As long as the client's future requests contains the cookie, and each request arrives within the stickytime interval, the client will maintain affinity with the initial server.
Active cookie affinity is used to ensure that a client continues to be load balanced to the same server for some period of time. This is accomplished by sending a cookie to be stored by the clients browser. The cookie contains the cluster:port that was used to make the decision, the server that was load balanced to, and a timeout timestamp for when the affinity is no longer valid. Whenever a rule fires that has active cookie affinity turned on, the cookie sent by the client is examined. If a cookie is found that contains the identifier for the cluster:port that fired, then the server load balanced to, and the expires timestamp are extracted from the cookie. If the server is still in the set used by the rule, and its weight is greater than zero, and the expires timestamp is greater than now, then the server in the cookie is chosen to load balance to. If any of the preceding three conditions are not met, a server is chosen using the normal algorithm. Once a server has been chosen (using either of the two methods) a new cookie is constructed containing IBMCBR, cluster:port:server_chosen information, and a timestamp. The timestamp will be the time that affinity expires. The "cluster:port:server_chosen" are encoded so that no information about the CBR configuration is revealed. An "expires" parameter is also inserted in the cookie. This parameter is in a format the browser can understand, and causes the cookie to become invalid two hours after the expires timestamp. This is so the client's cookie database isn't cluttered up.
This new cookie is then inserted in the headers that go back to the client, and if the client's browser is configured to accept cookies, it will send back subsequent requests.
The active cookie affinity option, for the rule command, can only be set to activecookie if port stickytime is zero (not enabled). Once active cookie affinity is active on a rule then you cannot enable stickytime on the port.
To enable active cookie affinity for a particular rule, use the rule set command:
rule set cluster:port:rule stickytime 60 rule set cluster:port:rule affinity activecookie
Making a rule sticky would normally be used for CGI or servlets that store client state on the server. The state is identified by a cookie ID (these are server cookies). Client state is only on the selected server, so the client needs the cookie from that server to maintain that state between requests.
Passive cookie affinity applies to the Dispatcher component's content-based routing (cbr) forwarding method and to the CBR component. See Dispatcher's content-based routing (cbr forwarding method) for information on how to configure Dispatcher's cbr forwarding method.
Passive cookie affinity provides a way to make clients sticky to a particular server. Once you enable the affinity of a rule to "passivecookie", passive cookie affinity allows you to load-balance Web traffic with affinity to the same server, based on self-identifying cookies generated by the servers. You configure passive cookie affinity at the rule level. Once the rule fires, if passive cookie affinity is enabled, Network Dispatcher will choose the server based on the cookie name in the HTTP header of the client request. Network Dispatcher will send new incoming requests to servers based on cookies that have been generated by servers during previous connections. If the cookie value in the client request is not found or does not match any of the servers' cookie values, the server will be chosen using the weighted round-robin technique.
To configure passive cookie affinity:
The passive cookie affinity option, for the rule command, can only be set to passivecookie if port stickytime is zero (not enabled). Once passive cookie affinity is active on a rule then you cannot enable stickytime on the port.
URI affinity applies to Dispatcher's cbr forwarding method and the CBR component. See Dispatcher's content-based routing (cbr forwarding method) for information on how to configure the cbr forwarding method.
URI affinity allows you to load-balance Web traffic to Caching Proxy servers which allow unique content to be cached on each individual server. As a result, you will effectively increase the size of your site's cache by eliminating redundant caching of content on multiple machines. Configure URI affinity at the rule level. Once the rule fires, if URI affinity is enabled and the same set of servers are up and responding, then Network Dispatcher will forward new incoming client requests with the same URI to the same server.
Typically, Network Dispatcher can distribute requests to multiple servers that serve identical content. When using Network Dispatcher with a group of caching servers, frequently accessed content eventually becomes cached on all the servers. This supports a very high client load by replicating identical cached content on multiple machines. This is particularly useful for high volume Web sites.
However, if your Web site supports a moderate volume of client traffic to very diverse content, and you prefer to have a larger cache spread across multiple servers, your site would perform better if each caching server contained unique content and Network Dispatcher distributed the request only to the caching server with that content.
With URI affinity, Network Dispatcher allows you to distribute the cached content to individual servers, eliminating redundant caching of content on multiple machines. Performance for diverse-content server sites using Caching Proxy servers will be improved with this enhancement. It will send identical requests to the same server, thereby caching content on single servers only. And, the effective size of the cache will grow larger with each new server machine added to the pool.
To configure URI affinity:
The URI affinity option, for the rule command, can only be set to URI if port stickytime is zero (not enabled). Once URI affinity is active on a rule then you cannot enable stickytime on the port.
This feature is only available for the Dispatcher component.
Dispatcher provides the ability to detect potential "denial of service" attacks and notify administrators via an alert. Dispatcher does this by analyzing incoming requests for a conspicuous amount of half-open TCP connections on servers, a common trait of simple denial of service attacks. In a denial of service attack, a site receives a large quantity of fabricated SYN packets from a large number of source IP addresses and source port numbers, but the site receives no subsequent packets for those TCP connections. This results in a large number of half-opened TCP connections on the servers, and over time the servers can become very slow, accepting no new incoming connections.
Network Dispatcher provides user exits that trigger scripts which you can customize that alert the Administrator to a possible denial of service attack. Dispatcher provides the following sample script files in the ...nd/servers/samples directory:
In order to run the files, you must move them to the ...nd/servers/bin directory and remove the ".sample" file extension.
To implement the DoS attack detection, set the maxhalfopen parameter on the ndcontrol port command as follows:
ndcontrol port set 127.40.56.1:80 maxhalfopen 1000
In the above example, Dispatcher will compare the current total number of half-open connections (for all servers residing on cluster 127.40.56.1 on port 80) with the threshold value of 1000 (specified by the maxhalfopen parameter). If the current half-open connections exceeds the threshold, then a call to an alert script (halfOpenAlert) is made. When the number of half-open connections drops below the threshold, a call to another alert script (halfOpenAlertDone) is made to indicate that the attack is over.
To determine how to set the maxhalfopen value: Periodically (perhaps every 10 minutes) run a half-open connection report (ndcontrol port halfopenaddressreport cluster:port) when your site is experiencing normal to heavy traffic. The half-open connection report will return the current "total half-open connections received." You should set maxhalfopen to a value that is anywhere from 50% to 200% larger than the largest number of half-open connnections that your site experiences.
In addition to statistical data reported, the halfopenaddressreport will also generate entries in the log (..nd/servers/logs/dispatcher/halfOpen.log) for all the client addresses (up to approximately 8000 address pairs) that have accessed servers that resulted in half open connnections.
To provide additional protection from denial of service attacks for backend servers, you can configure wildcard clusters and ports. Specifically, under each configured cluster add a wildcard port with no servers. Also add a wildcard cluster with a wildcard port and no servers. This will have the effect of discarding all packets which are not addressed to a non-wildcard cluster and port. For information on wildcard clusters and wildcard ports, see Use wildcard cluster to combine server configurations and Use wildcard port to direct unconfigured port traffic.
The binary logging feature allows server information to be stored in binary files. These files can then be processed to analyze the server information that has been gathered over time.
The following information is stored in the binary log for each server defined in the configuration.
Some of this information is retrieved from the executor as part of the manager cycle. Therefore the manager must be running in order for the information to be logged to the binary logs.
Use ndcontrol log command set to configure binary logging.
The start option starts logging server information to binary logs in the logs directory. One log is created at the start of every hour with the date and time as the name of the file.
The stop option stops logging server information to the binary logs. The log service is stopped by default.
The set interval option controls how often information is written to the logs. The manager will send server information to the log server every manager interval. The information will be written to the logs only if the specified log interval seconds have elapsed since the last record was written to the log. By default, the log interval is set to 60 seconds. There is some interaction between the settings of the manger interval and the log interval. Since the log server will be provided with information no faster than manager interval seconds setting the log interval less than the manager interval effectively sets it to the same as the manager interval. This logging technique allows you to capture server information at any granularity. You can capture all changes to server information that are seen by the manager for calculating server weights. However, this amount of information is probably not required to analyze server usage and trends. Logging server information every 60 seconds gives you snapshots of server information over time. Setting the log interval very low can generate huge amounts of data.
The set retention option controls how long log files are kept. Log files older than the retention hours specified will be deleted by the log server. This will only occur if the log server is being called by the manager, so stopping the manager will cause old log files not to be deleted.
The status option returns the current settings of the log service. These settings are whether the service is started, what the interval is, and what the retention hours are.
A sample Java program and command file have been provided in the ...nd/servers/samples/BinaryLog directory. This sample shows how to retrieve all the information from the log files and print it to the screen. It can be customized to do any type of analysis you want with the data. An example using the supplied script and program for the dispatcher would be:
ndlogreport 2001/05/01 8:00 2001/05/01 17:00
to get a report of the Dispatcher component's server information from 8:00 AM to 5:00 PM on May 1, 2001. (For CBR, use cbrlogreport. For Mailbox Locator, use mllogreport. For Cisco Consultant use lbclogreport.)
In Cisco Consultant, the Cisco CSS Switch performs the tasks done by the executor in the Dispatcher component. Along with the current weight for each server and some other information required for its calculations, the manager gets the active and new connections values from the Cisco CSS Switch. These values are based on information that is generated and stored internally in the Cisco CSS Switch.
Cisco Consultant queries the Cisco CSS Switch management information base (MIB) to obtain active and new connection information and receives the following:
apSvcConnections OBJECT-TYPE SYNTAX Integer32 MAX-ACCESS read-only STATUS current DESCRIPTION "The current number of TCP connections to this service" DEFVAL { 0 } --DEFAULT ap-display-name Service Connections ::= {apSvcEntry 20}
The apSvcConnections object identifier is:
1.3.6.1.4.1.2467.1.15.2.1.20
The number of active connections is dependent upon the number of clients, as well as the length of time necessary to use the services that are provided by the load-balanced server machines. If the client connections are fast (such as small Web pages served using HTTP GET), the number of active connections is fairly low. If the client connections are slower (such as a database query), then the number of active connections is higher.
The index for this variable is:
INDEX { apCntsvcOwnName, apCntsvcCntName, apCntsvcSvcName }Following is the MIB entry.
apCntsvcHits OBJECT-TYPE SYNTAX Integer32 MAX-ACCESS read-only STATUS current DESCRIPTION "Total number of flows placed onto this service for this content rule." DEFVAL { 0 } --DEFAULT ap-display-name Hits --DEFAULT apjam-popup-ref apCntSvcInst, Statistics --DEFAULT apjam-chart-def cntSvcHitsChart, pie, apCntInst, "Hit Information Per Service: --DEFAULT apjam-chart-item cntSvcHitsChart, getnext, apCntsvcSvcName ::= {apSvcEntry 20}
The apCntsvcHits object identifier is:
1.3.6.1.4.1.2467.1.18.2.1.4
The Cisco CSS Switch must be configured to use weighted round-robin load balancing. Refer to "Configuring Weight" in the Content Services Switch Basic Configuration Guide for information on how to do this.
Weights are set by the manager function based on internal counters in the Cisco CSS Switch and feedback from the advisors and Metric Server. If you want to set weights manually while running the manager, specify the fixedweight option on the lbccontrol server command
If all servers are down, all weights are zero. In a case such as this, when no servers are processing requests because all weights are zero, weights are set to one-half the weightbound to allow an equal chance of request processing from any capable server. The monitor shows the true weight values of zero; however, Cisco Consultant displays a weight of one-half weightbound in all other places.
Weights are sent to the Cisco CSS Switch using SNMP. Cisco Consultant sets apSvcWeight in svcExt.mib. Following is the apSvcWeight entry.
apSvcWeight OBJECT-TYPE SYNTAX Integer 32(1..10) MAX-ACCESS read-create STATUS current DESCRIPTION "The service weight which is used in conjunction with load metrics when making load allocation decisions. The weight may be used to bias flows towards the specified service." DEFVAL { 1 } --DEFAULT ap-display-name Service Weight --DEFAULT apjam-popup-ref apServicesGroupInst, Properties, Advanced --DEFAULT apjam-wizard-field 2, normal ::= {apSvcEntry 16}
The apSvcWeight object identifier is:
1.3.6.1.4.1.2467.1.15.2.1.12
Weights are applied to all servers on a port. For any particular port, the requests are distributed between servers based on their weights relative to each other. For example, if one server is set to a weight of 10, and the other to 5, the server set to 10 should get twice as many requests as the server set to 5.
To specify the maximum weight boundary that any server can have, use the lbccontrol port set weightbound command. This command specifies the differences in the number of requests each server gets. If you set the maximum weight to 1, all the servers can have a weight of 1, 0 if suspended, or -1 if marked down. As you increase this number, the difference in how servers are weighted is increased. At a maximum weight of 2, one server can get twice as many requests as another.
If an advisor learns that a server is offline, it tells the manager, and the manager sets the weight for the server to zero. When the weight of a server is greater than zero, the weight is sent to the Cisco CSS Switch, and the server becomes active; however, if the server weight is less than or equal to zero, the server is suspended. Activating and suspending a service is accomplished by setting the apSvcEnable MIB variable in the Cisco CSS Switch svcExt.mib. Following is the apSvcEnable MIB entry.
apSvcEnable OBJECT-TYPE SYNTAX Integer disable(0) enable(1) MAX-ACCESS read-create STATUS current DESCRIPTION "The state of the service, either enabled or disabled." DEFVAL { disable } --DEFAULT ap-display-name Status --DEFAULT apjam-popup-ref apServicesGroupInst, Properties --DEFAULT apjam-wizard-field 2, normal ::= {apSvcEntry 12}
The apSvcEnable object identifier is:
1.3.6.1.4.1.2467.1.15.2.1.16
This chapter explains how to operate and manage Network Dispatcher and includes the following sections:
Network Dispatcher provides an option to run its configuration programs on a machine other than the one running the Network Dispatcher servers.
Communication between the configuration programs (ndcontrol, cbrcontrol, mlcontrol, sscontrol, lbccontrol, ndwizard, cbrwizard, mlwizard, sswizard, ndadmin) is performed using Java Remote Method Invocation (RMI) calls. The command to connect to a Network Dispatcher machine for remote administration is ndcontrol host:remote_host. If the RMI call comes from a machine other than the local machine, a public key/private key authentication sequence must occur before the configuration command will be accepted.
Communication between the control programs running on the same machine as the component servers are not authenticated.
Use the following command to generate public and private keys to be used for remote authentication:
This command runs only on the same machine as the Network Dispatcher.
Using the create option creates a public key in the servers key directory (...nd/servers/key/) and creates private keys in the administration keys directory (...nd/admin/keys/) for each of the Network Dispatcher components. The file name for the private key is: component-ServerAddress-RMIport. These private keys must then be transported to the remote clients and placed in the administration keys directory.
For a Network Dispatcher machine with hostname address 10.0.0.25 using the default RMI port for each component, the ndkeys create command generates the following files:
The administration fileset has been installed on another machine. The private key files must be placed in .../nd/admin/keys directory on the remote client machine.
The remote client will now be authorized to configure Network Dispatcher on 10.0.0.25.
These same keys must be used on all remote clients that you want to authorize to configure Network Dispatcher on 10.0.0.25.
If you were to run the ndkeys create command again, a new set of public/private keys would be generated. This would mean that all remote clients who tried to connect using the previous keys would not be authorized. The new key would have to be placed in the correct directory on those clients you want to reauthorize.
The ndkeys delete command deletes the public and private keys on the server machine. If these keys are deleted, no remote clients will be authorized to connect to the servers.
For both ndkeys create and ndkeys delete there is a force option. The force option suppresses the command prompts that ask if you wish to overwrite or delete the existing keys.
Network Dispatcher posts entries to a server log, a manager log, a metric monitor log (logging communications with Metric Server agents), and a log for each advisor you use.
You can set the logging level to define the expansiveness of the messages written to the log. At level 0, errors are logged and Network Dispatcher also logs headers and records of events that happen only once (for example, a message about an advisor starting to be written to the manager log). Level 1 includes ongoing information, and so on, with level 5 including every message produced to aid in debugging a problem if necessary. The default for the server log is 0. The default for the manager, advisor, and subagent logs is 1.
You can also set the maximum size of a log. When you set a maximum size for the log file, the file will wrap; when the file reaches the specified size, the subsequent entries will be written at the top of the file, overwriting the previous log entries. You cannot set the log size to a value that is smaller than the current one. Log entries are timestamped so you can tell the order in which they were written.
The higher you set the log level, the more carefully you should choose the log size. At level 0, it is probably safe to leave the log size to the default of 1MB; however, when logging at level 3 and above, you should limit the size without making it too small to be useful.
By default, the logs generated by the Network Dispatcher will be stored in the logs directory of the Network Dispatcher installation. To change this path, set the nd_logdir variable in the ndserver script.
AIX, Linux, and Solaris: The ndserver script is found in /usr/bin directory. In this script, the variable nd_logdir is set to the default directory. You can modify this variable to specify your log directory. Example:
ND_LOGDIR=/path/to/my/logs/
Windows 2000: The ndserver file is found in the Windows 2000 system directory, typically C:\WINNT\SYSTEM32. In the ndserver file, the variable nd_logdir is set to the default directory. You can modify this variable to specify your log directory. Example:
set ND_LOGDIR=c:\path\to\my\logs\
For all operating systems, make sure that there are no spaces on either side of the equal sign and that the path ends in a slash ("/" or "\" as appropriate).
The binary logging feature of Network Dispatcher uses the same log directory as the other log files. See Using binary logging to analyze server statistics.
This section explains how to operate and manage the Dispatcher component.
For Network Dispatcher, connections are considered stale when there has been no activity on that connection for the number of seconds specified in stale timeout. When the number of seconds has been exceeded with no activity, Network Dispatcher will remove that connection record from its tables, and subsequent traffic for that connection will be discarded.
At the port level, for example, you can specify the stale timeout value on the ndcontrol port set staletimeout command.
Stale timeout can be set at the executor, cluster, and port levels. At the executor and cluster levels, the default is 300 seconds and it filters down to the port. At the port level, the default depends on the port. Some well defined ports have different default stale timeout values. For example, the telnet port 23 has a default of 32,000,000 seconds.
Some services may also have staletimeout values of their own. For example, LDAP (Lightweight Directory Access Protocol) has a configuration parameter called idletimeout. When idletimeout seconds have been exceeded, an idle client connection will be forcibly closed. Idletimeout may also be set to 0, which means that the connection will never be forcibly closed.
Connectivity problems can occur when Network Dispatcher's stale timeout value is smaller than the service's timeout value. In the case of LDAP, the Network Dispatcher staletimeout value defaults to 300 seconds. If there is no activity on the connection for 300 seconds, Network Dispatcher will remove the connection record from its tables. If the idletimeout value is larger than 300 seconds (or set to 0), the client may still believe that it has a connection to the server. When the client sends packets, the packets will be discarded by Network Dispatcher. This will cause LDAP to hang when a request is made to the server. To avoid this problem, set the LDAP idletimeout to a nonzero value that is the same or smaller than the Network Dispatcher staletimeout value.
A client sends a FIN packet after it has sent all its packets so that the server will know that the transaction is finished. When Dispatcher receives the FIN packet, it marks the transaction from active state to FIN state. When a transaction is marked FIN, the memory reserved for the connection can be cleared by the garbage collector built into the executor.
You can use the FIN timeout and count to set how often the executor will perform garbage collection and how much it will perform. The executor periodically checks the list of connections it has allocated. When the number of connections in the FIN state is greater than or equal to the FIN count, the executor will attempt to free the memory used to hold this connection information. You can change the FIN count by entering the ndcontrol executor set fincount command.
The garbage collector frees the memory for any connection that is in the FIN state and is older than the number of seconds specified in the FIN timeout. You can change the FIN timeout by entering the ndcontrol executor set fintimeout command.
Stale timeout value is the number of seconds during which there can be no activity on a connection before that connection is removed. See Using stale timeout value, for more information. The FIN count also affects how often "stale" connections are removed. If you have little memory on your Dispatcher machine, you should set the FIN count lower. If you have a busy Web site, you should set the FIN count higher.
Various charts can be displayed based on information from the executor and relayed to the manager. (The GUI Monitor menu option requires that the manager function is running):
A network management system is a program that runs continuously and is used to monitor, reflect status of, and control a network. Simple Network Management Protocol (SNMP), a popular protocol for communicating with devices in a network, is the current network management standard. The network devices typically have an SNMP agent and one or more subagents. The SNMP agent talks to the network management station or responds to command line SNMP requests. The SNMP subagent retrieves and updates data and gives that data to the SNMP agent to communicate back to the requester.
Dispatcher provides an SNMP Management Information Base (ibmNetDispatcherMIB) and an SNMP subagent. This allows you to use any network management system, such as -- Tivoli NetView, Tivoli Distributed Monitoring, or HP OpenView -- to monitor the Dispatcher's health, throughput, and activity. The MIB data describes the Dispatcher being managed and reflects current Dispatcher status. The MIB gets installed in the ..nd/admin/MIB subdirectory.
The network management system uses SNMP GET commands to look at MIB values on other machines. It then can notify you if specified threshold values are exceeded. You can then affect Dispatcher performance, by modifying configuration data for Dispatcher, to proactively tune or fix Dispatcher problems before they become Dispatcher or Web server outages.
The system usually provides an SNMP agent for each network management station. The user sends a GET command to the SNMP agent. In turn, this SNMP agent sends a GET command to retrieve the specified MIB variable values from a subagent responsible for those MIB variables.
Dispatcher provides a subagent that updates and retrieves MIB data. The subagent responds with the appropriate MIB data when the SNMP agent sends a GET command. The SNMP agent communicates the data to the network management station. The network management station can notify you if specified threshold values are exceeded.
The Dispatcher SNMP support includes an SNMP subagent that uses Distributed Protocol Interface (DPI) capability. DPI is an interface between an SNMP agent and its subagents.
Figure 28. SNMP commands for AIX and Solaris
AIX provides an SNMP agent that uses SNMP Multiplexer protocol (SMUX) and provides DPID2, which is an additional executable that works as a translator between DPI and SMUX.
For Solaris, you must obtain an SNMP agent that is SMUX-enabled since Solaris does not provide one. Network Dispatcher provides DPID2 for Solaris in the /opt/nd/servers/samples/SNMP directory.
The DPI agent must run as a root user. Before you execute the DPID2 daemon, update the /etc/snmpd.peers file and the /etc/snmpd.conf file as follows:
"dpid2" 1.3.6.1.4.1.2.3.1.2.2.1.1.2 "dpid_password"
smux 1.3.6.1.4.1.2.3.1.2.2.1.1.2 dpid_password #dpid
Refresh snmpd so that it will reread the /etc/snmpd.conf file:
refresh -s snmpd
Start the DPID SMUX peer:
dpid2
The daemons must be started in the following order:
Figure 29. SNMP commands for Windows 2000
To get a DPI-capable SNMP agent for Windows 2000, install the Windows NT version of the IBM SystemView Agent toolkit from http://www.tivoli.com/support/sva.
Before you start the SystemView SNMPD process, you must disable the Microsoft Windows SNMP support. The SystemView snmpd supports DPI subagents and Microsoft-compliant agents.
To disable the Windows SNMP support:
To configure the SystemView SNMP agent, follow the instructions in Providing a community name for SNMP.
You should configure the SNMP community name. The default SNMP community name is public. In UNIX systems, this is set up in a file named /etc/snmpd.conf.
On all systems, the community name must be configured and used consistently. That is, if the community name for Network Dispatcher is set to "OurCommunity" in the SNMP agent configuration, it must also be set to "OurCommunity" in the subagent configuration.
For Windows 2000, before creating community name, configure the IBM SystemView SNMP agent.
This step allows any host in any network to access the SNMP MIB variables. After you have verified that these values work, you can change them according to your requirements.
With the executor running, use the ndcontrol subagent start [communityname] command to define the community name used between the Dispatcher DPI subagent and the SNMP agent. The default for community name is public. If you change this value, you must also add the new community name to the SystemView Agent using snmpcfg as above.
SNMP communicates by sending and receiving traps, messages sent by managed devices to report exception conditions or the occurrence of significant events, such as a threshold having been reached.
The subagent uses the following traps:
The indHighAvailStatus trap announces that the value of the high-availability status state variable (hasState) has changed. The possible values of hasState are:
The indSrvrGoneDown trap announces that the weight for the server specified by the csAddr, psNum, ssAddr portion of the Object Identifier has gone to zero. The last known number of active connections for the server is sent in the trap. This trap indicates that, as far as the Dispatcher can determine, the specified server has gone down.
The indDOSAttack trap indicates that numhalfopen, the number of half-open connections consisting only of SYN packets, has exceeded the maxhhalfopen threshold for the port specified by the csAddr, psNum portion of the Object Identifier. The number of servers configured on the port is sent in the trap. This trap indicates that Network Dispatcher may be experiencing a Denial Of Service Attack.
The indDOSAttackDone trap indicates that numhalfopen, the number of half-open connections consisting only of SYN packets, has fallen below the maxhalfopen threshold for the port specified by the csAddr, psNum portion of the Object Identifier. The number of servers configured on the port is sent in the trap. When Network Dispatcher determines that the possible Denial of Service attack is over, this trap will be sent after an indDOSAttack trap is sent.
Due to a limitation in the SMUX API, the enterprise identifier reported in traps from the ibmNetDispatcher subagent may be the enterprise identifier of dpid2, instead of the enterprise identifier of ibmNetDispatcher, 1.3.6.1.4.1.2.6.144. However, the SNMP management utilities will be able to determine the source of the trap because the data will contain an object identifier from within the ibmNetDispatcher MIB.
The ndcontrol subagent start command turns the SNMP support on. The ndcontrol subagent stop command turns the SNMP support off.
For more information about the ndcontrol command, see ndcontrol subagent -- configure SNMP subagent.
Built into the Linux kernel is a firewall facility called ipchains. When Network Dispatcher and ipchains run concurrently, Network Dispatcher sees packets first, followed by ipchains. This allows the use of ipchains to harden a Linux Network Dispatcher box, which could be, for example, a Network Dispatcher box that is used to load balance firewalls.
When ipchains or iptables is configured as completely restricted (no inbound or outbound traffic permitted), the packet-forwarding portion of Network Dispatcher continues to function normally.
Note that ipchains and iptables cannot be used to filter incoming traffic before it is load balanced.
Some additional traffic must be permitted for all of Network Dispatcher to function properly. Some examples of this communication are:
In general, an appropriate ipchains strategy for the Network Dispatcher boxes is to disallow all traffic, except that which is to or from the back-end servers, the partner high availability Network Dispatcher, any reach targets, or any configuration hosts.
This section explains how to operate and manage the CBR component of Network Dispatcher.
CBR and Caching Proxy collaborate via the Caching Proxy plugin API to handle HTTP and HTTPS (SSL) request. Caching Proxy must be running on the same machine in order for CBR to begin load balancing servers. Set up CBR and Caching Proxy as described in CBR configuration example.
After starting CBR, you can control it using either of the following methods:
The logs used by CBR are similar to those used in Dispatcher. For more information, see Using Network Dispatcher logs.
After starting Mailbox Locator, you can control it using either of the following methods:
The logs used by Mailbox Locator are similar to those used in Dispatcher. For more description, see Using Network Dispatcher logs.
After starting Site Selector, you can control it using either of the following methods:
The logs used by Site Selector are similar to those used in Dispatcher. For more description, see Using Network Dispatcher logs.
After starting Cisco Consultant, you can control it using either of the following methods:
The logs used by Cisco Consultant are similar to those used in Dispatcher. For more description, see Using Network Dispatcher logs.
Metric Server provides server load information to the Network Dispatcher. Metric Server resides on each of the servers that are being load balanced.
Change the log level in the Metric Server startup script. You can specify a log level range of 0 through 5, similar to the log level range in Network Dispatcher logs. This will generate an agent log in the ...ms/logs directory.
This chapter helps you detect and resolve problems associated with Network Dispatcher. Find the symptom you are experiencing in Troubleshooting tables.
These are the troubleshooting tables for Dispatcher, CBR, Mailbox
Locator, Site Selector, and Consultant for Cisco CSS Switches.
Table 14. Dispatcher troubleshooting table
Symptom | Possible Cause | Go to... |
---|---|---|
Dispatcher not running correctly | Conflicting port numbers | Checking Dispatcher port numbers |
Configured a collocated server and it will not respond to load balanced requests | Wrong or conflicting address | Problem: Dispatcher and server will not respond |
Connections from client machines not being served or connections timing out |
| Problem: Dispatcher requests are not being balanced |
Client machines are not being served or are timing out | High availability not working | Problem: Dispatcher high-availability function is not working |
Unable to add heartbeat (Windows 2000) | Source address is not configured on an adapter | Problem: Unable to add heartbeat (Windows 2000) |
Server not serving requests (Window) | An extra route has been created in the routing table | Problem: Extra routes (Windows 2000) |
Advisors not working correctly with wide area | Advisors are not running on remote machines | Problem: Advisors not working correctly |
SNMPD will not start or will not continue to run (Windows 2000) | The community name passed in the SNMP commands does not agree with the community name with which the subagent was started | Problem: SNMPD does not run correctly (Windows 2000) |
Dispatcher, Microsoft IIS, and SSL are not working or will not continue | Unable to send encrypted data across protocols | Problem: Dispatcher, Microsoft IIS, and SSL do not work (Windows 2000) |
Connection to remote machine refused | Older version of the keys is still being used | Problem: Dispatcher connection to a remote machine |
The ndcontrol or ndadmin command fails with 'Server not responding' or 'unable to access RMI server' message |
| Problem: ndcontrol or ndadmin command fails |
"Cannot Find the File..." error message, when running Netscape as default browser to view online help (Windows 2000) | Incorrect setting for HTML file association | Problem: "Cannot find the file... error message when trying to view online Help (Windows 2000) |
"stty: : No such device or address" error message, when starting ndserver on Solaris 2.7. | Please disregard this error message. This is not a problem. Ndserver will run correctly | Problem: Spurious error message when starting ndserver on Solaris 2.7 |
Graphical user interface does not start correctly | Insufficient paging space | Problem: Graphical user interface (GUI) does not start correctly |
Error running Dispatcher with Caching Proxy installed | Caching Proxy file dependency | Problem: Error running Dispatcher with Caching Proxy installed |
Graphical user interface does not display correctly. | Resolution is incorrect. | Problem: Graphical user interface (GUI) does not display correctly |
Help panels sometimes disappear behind other windows | Java limitation | Problem: On Windows 2000, help windows sometimes disappear behind other open windows |
Network Dispatcher cannot process and forward a frame | Need a unique MAC address for each NIC | Problem: Network Dispatcher cannot process and forward a frame |
Blue screen appears | No installed and configured network card | Problem: A blue screen displays when you start the Network Dispatcher executor |
Path to Discovery prevents return traffic | The cluster is aliased on the loopback | Problem: Path to Discovery prevents return traffic with Network Dispatcher |
Advisors show that all servers are down | TCP checksum is not computed correctly | Problem: Advisors show that all servers are down |
High availability in the Wide Area mode of Network Dispatcher does not work. | Remote Dispatcher must be defined as a server in a cluster on local Dispatcher | Problem: High availability in the Wide Area mode of Network Dispatcher does not work |
GUI hangs (or unexpected behavior) when trying to load a large configuration file. | Java does not have access to enough memory to handle such a large change to the GUI. | Problem: GUI hangs (or unexpected behavior) when trying to load a large configuration file |
Table 15. CBR Troubleshooting table
Symptom | Possible Cause | Go to... |
CBR not running correctly | Conflicting port numbers | Checking CBR port numbers |
The cbrcontrol or ndadmin command fails with 'Server not responding' or 'unable to access RMI server' message | Commands fail due to socksified stack. Or commands fail due to not starting cbrserver | Problem: cbrcontrol or ndadmin command fails |
Requests are not being load balanced | Caching Proxy was started before the executor was started | Problem: Requests not being load balanced |
On Solaris, the cbrcontrol executor start command fails with 'Error: Executor was not started.' message | Command fails because the system IPC defaults may need to be modified | Problem: On Solaris, cbrcontrol executor start command fails |
URL rule doesn't work | Syntactical or configuration error | Problem: Syntactical or configuration error |
Table 16. Mailbox Locator Troubleshooting table
Symptom | Possible Cause | Go to... |
Mailbox Locator not running correctly | Conflicting port numbers | Checking Mailbox Locator port numbers |
The mlserver command returns a "java.rmi.RMI Security Exception: security.fd.read" exception | The system's limit on file descriptors is too small for the number of requests that mlserver is trying to service | Problem: The mlserver command is stopped |
The mlcontrol or ndadmin command fails with 'Server not responding' or 'unable to access RMI server' message | Commands fail due to socksified stack. Or commands fail due to not starting mlserver. | Problem: mlcontrol or ndadmin command fails |
Unable to add a port | Another application is already listening to that port | Problem: Unable to add a port |
Receive proxy error when trying to add a port | The cluster address was not configured on a NIC before the proxy was started. Or, another application is running on that port. | Problem: Receive proxy error when trying to add a port |
Table 17. Site Selector troubleshooting table
Symptom | Possible Cause | Go to... |
---|---|---|
Site Selector not running correctly | Conflicting port number | Checking Site Selector port numbers |
Site Selector does not round-robin incoming requests from Solaris client | Solaris systems run a "name service cache daemon" | Problem: Site Selector doesn't round-robin traffic from Solaris clients |
The sscontrol or ndadmin command fails with 'Server not responding' or 'unable to access RMI server' message | Commands fail due to socksified stack. Or commands fail due to not starting ssserver. | Problem: sscontrol or ndadmin command fails |
ssserver fails to start on Windows 2000 | Windows does not require the host name to be in the DNS. | Problem: ssserver is failing to start on Windows 2000 |
Machine with duplicate routes not load balancing correctly -- name resolution appears to fail | Site Selector machine with multiple adapters attached to the same subnet | Problem: Site Selector with duplicate routes not load balancing correctly |
Table 18. Consultant for Cisco CSS Switches troubleshooting table
Symptom | Possible Cause | Go to... |
---|---|---|
lbcserver will not start | Conflicting port numbers | Checking Cisco Consultant port numbers |
The lbccontrol or ndadmin command fails with 'Server not responding' or 'unable to access RMI server' message | Commands fail due to socksified stack. Or commands fail due to not starting lbcserver. | Problem: lbccontrol or ndadmin command fails |
receive error: Cannot create registry on port 14099 | Expired product license | Problem: Cannot create registry on port 14099 |
Table 19. Metric Server troubleshooting table
Symptom | Possible Cause | Go to... |
---|---|---|
Metric Server IOException on Windows 2000 running .bat or .cmd user metric files | Full metric name is required | Problem: Metric Server IOException on Windows 2000 running .bat or .cmd user metric files |
Metric Server not reporting the load information to the Network Dispatcher machine | Possible causes include:
| Problem: Metric Server not reporting loads to Network Dispatcher machine |
Metric Server log reports "Signature is necessary for access to agent" when key files transferred to server | Key file fails authorization due to corruption. | Problem: Metric Server log reports Signature is necessary for access to agent |
If you are experiencing problems running the Dispatcher, it may be that one of your applications is using a port number that the Dispatcher normally uses. Be aware that the Dispatcher server uses the following port numbers:
If another application is using one of the Dispatcher port numbers, you can change the Dispatcher's port number by doing the following:
If you are experiencing problems running the CBR, it may be that one of your applications is using a port number that CBR normally uses. Be aware that CBR uses the following port number:
If another application is using one of the CBR port numbers, you can change the CBR's port number by doing the following:
If you are experiencing problems running the Mailbox Locator, it may be that one of your applications is using a port number that Mailbox Locator normally uses. Be aware that Mailbox Locator uses the following port numbers:
If another application is using one of the Mailbox Locator port numbers, you can change the Mailbox Locator's port number by doing the following:
If you are experiencing problems running the Site Selector component, it may be that one of your applications is using a port number that Site Selector normally uses. Be aware that Site Selector uses the following port numbers:
If another application is using one of the Site Selector port numbers, you can change the Site Selector's port number by doing the following:
If you are experiencing problems running the Cisco Consultant component, it may be that another application is using one of the port numbers used by Cisco Consultant's lbcserver. Be aware that Cisco Consultant uses the following port numbers:
14099 to receive commands from lbccontrol
10004 to send metric queries to Metric Server
If another application is using one of the Consultant port numbers, you can change the port numbers for Consultant by doing the following:
This problem can occur when another application is using one of the ports used by the Dispatcher. For more information, go to Checking Dispatcher port numbers.
This problem occurs when another address is being used other than the address specified. When collocating the Dispatcher and server, be sure that the server address used in the configuration is the NFA address or is configured as collocated.
This problem has symptoms such as connections from client machines not being served or connections timing out. Check the following to diagnose this problem:
This problem appears when a Dispatcher high-availability environment is configured and connections from the client machines are not being served or are timing out. Check the following to correct or diagnose the problem:
This Windows 2000 error occurs when the source address is not configured on an adapter. Check the following to correct or diagnose the problem.
ndconfig tr0 <ip address> netmask <netmask> or ndcontrol cluster configure
After setting up server machines, you may find that you have inadvertently created one or more extra routes. If not removed, these extra routes will prevent the Dispatcher from operating. To check for and delete them, see Setting up server machines for load balancing.
If you are using wide area support, and your advisors do not seem to be working correctly, make sure that they are started on both the local and the remote Dispatchers. See Using remote advisors with wide area support.
When using SNMP subagents, if the SystemView Agent SNMP daemon does not start and stay up, be sure that you have configured your SNMP community using the snmpcfg program. To access SNMP data from the Dispatcher subagent, the community name passed in the SNMP commands must agree with the community name with which the subagent was started.
When using Dispatcher, Microsoft IIS, and SSL, if they do not work together, there may be a problem with enabling SSL security. For more information about generating a key pair, acquiring a certificate, installing a certificate with a key pair, and configuring a directory to require SSL, see the Microsoft Information and Peer Web Services Information and Planning Guide, which comes with Windows 2000. The local URL for the document, which is viewed by a web browser, is: file:///C|/WINNT/system32/inetsrv/iisadmin/htmldocs/inetdocs.htm.
Dispatcher uses keys to allow you to connect to a remote machine and configure it. The keys specify an RMI port for the connection. It is possible to change the RMI port for security reasons or conflicts. When you change the RMI ports, the filename of the key is different. If you have more than one key in your keys directory for the same remote machine, and they specify different RMI ports, the command line will only try the first one it finds. If it is the incorrect one, the connection will be refused. The connection will not occur unless you delete the incorrect key.
EXCLUDE-MODULE java EXCLUDE-MODULE jre EXCLUDE-MODULE jrew EXCLUDE-MODULE javaw
The random port can cause problems when one of the administration consoles runs on the same machine as a firewall or through a firewall. For example, when Network Dispatcher runs on the same machine as a firewall, and you issue ndcontrol commands, you might see errors such as Error: Server not responding.
To avoid this problem, edit the ndserver script file (located in the PATH) to set the random port used by RMI. Include -DND_RMI_SERVER_PORT=yourPort within the END_ACCESS string, where yourPort is the port you specify.
For example:
END_ACCESS='-DND_CLIENT_KEYS_DIRECTORY=/usr/lpp/nd/admin/keys/dispatcher -DND_SERVER_KEYS_DIRECTORY=/usr/lpp/nd/dispatcher/key -DND_RMI_SERVER_PORT=10100' ND_RMIPORT=10099
Once complete, restart ndserver and open traffic for ports 10099 and 10100, or for the chosen port for the host address from which the administration console will be run.
For Windows 2000 when using Netscape as your default browser, the error message which results with this problem is: "Cannot find the file '<filename>.html' (or one of its components). Make sure the path and filename are correct and that all required libraries are available."
The problem is due to an incorrect setting for HTML file association. The solution is the following:
When starting ndserver on Solaris 2.7 platforms, the following spurious error message appears: "stty: : No such device or address." Please disregard this error message. Ndserver will run correctly.
The graphical user interface (GUI), which is ndadmin, requires a sufficient amount of paging space to function correctly. If insufficient paging space is available, the GUI might not start up completely. If this occurs, check your paging space and increase it if necessary.
If you uninstall Network Dispatcher to reinstall another version and get an error when you attempt to start the Dispatcher component, check to see if Caching Proxy is installed. Caching Proxy has a dependency on one of the Dispatcher files; this file will uninstall only when Caching Proxy is uninstalled.
To avoid this problem:
If you experience a problem with the appearance of the Network Dispatcher GUI, check the setting for the operating system's desktop resolution. The GUI is best viewed at a resolution of 1024x768 pixels.
When you first open help windows on Windows 2000, they sometimes disappear into the background behind existing windows. If this occurs, click on the window to bring it forward again.
On Solaris each network adapter has the same MAC address by default. This works properly when each adapter is on a different IP subnet; however, in a switched environment, when multiple NICs with the same MAC and the same IP subnet address communicate with the same switch, the switch sends all traffic bound for the single MAC (and both IPs) down the same wire. Only the adapter that last put a frame on the wire sees the IP packets bound for both adapters. Solaris might discard packets for a valid IP address that arrived on the "wrong" interface.
If all network interfaces are not designated for Network Dispatcher as configured in ibmnd.conf, and if the NIC that is not defined in ibmnd.conf receives a frame, Network Dispatcher does not have the ability to process and forward the frame.
To avoid this problem, you must override the default and set a unique MAC address for each interface. Use this command:
ifconfig interface ether macAddr
For example:
ifconfig hme0 ether 01:02:03:04:05:06
On Windows 2000, you must have a network card installed and configured before starting the executor.
The AIX operating system contains a networking parameter called path MTU discovery. During a transaction with a client, if the operating system determines that it must use a smaller maximum transmission unit (MTU) for the outgoing packets, path MTU discovery has AIX create a route to remember that data. The new route is for that specific client IP and records the necessary MTU to reach it.
When the route is being created, a problem might occur on the servers resulting from the cluster being aliased on the loopback. If the gateway address for the route falls in the subnet of the cluster/netmask, AIX creates the route on the loopback. This happens because that was the last interface aliased with that subnet.
For example, if the cluster is 9.37.54.69 and a 255.255.255.0 netmask is used, and the intended gateway is 9.37.54.1, AIX uses the loopback for the route. This causes the server's responses to never leave the box, and the client times out waiting. The client typically sees one response from the cluster, then the route is created and the client receives nothing more.
There are two solutions to this problem.
Notes:
Windows 2000 has a new feature called Task Offload that allows the TCP checksum to be calculated by the adapter card rather than the operating system. This improves performance on the system. If Task Offload is enabled, Network Dispatcher advisors report that servers are down when they are not.
The problem is that the TCP checksum is not computed correctly for packets coming from the cluster address, which is what happens with advisor traffic.
To avoid this problem, go to the adapter card settings and disable Task Offload.
This problem was first observed with Adaptec's ANA62044 QuadPort Adapter. This adapter card refers to the feature as the Transmit Checksum offload. Disable Transmit Checksum offload to avoid the problem.
When you set up a Wide Area Network Dispatcher, you must define the remote Dispatcher as a server in a cluster on your local Dispatcher. Typically, you use the non-forwarding address (NFA) of the remote Dispatcher as the destination address of the remote server. If you do this, and then set up high availability on the remote Dispatcher, it will fail. This happens because the local Dispatcher always points to the primary on the remote side when you use its NFA to access it.
To get around this problem:
When the remote primary Dispatcher comes up, it will alias this address on its adapter, allowing it to accept traffic. If a failure occurs, the address moves to the backup machine and the backup continues to accept traffic for that address.
When trying to load a large configuration file (roughly 200 or more add commands), the GUI may hang or display unexpected behavior, such as responding to screen changes at an extremely slow rate of speed.
This occurs because Java does not have access to enough memory to handle such a large change to the GUI.
There is an option on the runtime environment that can be specified to increase the memory allocation pool available to Java.
The option is -Xmxn where n is the maximum size, in bytes, for the memory allocation pool. n must be a multiple of 1024 and must be greater than 2MB. The value n may be followed by k or K to indicate kilobytes, or m or M to indicate megabytes. For example, -Xmx128M and -Xmx81920k are both valid. The default value is 64MB. Solaris 7 and Solaris 8 SPARC platforms have a maximum value of 4000m; Solaris 2.6 and x86 platforms have a maximum value of 2000m.
To add this option, modify the ndadmin script file as follows:
START jrew -mx64m %END_ACCESS% %CONFIG_DIR% -DEND_INSTALL_PATH=%IBMNDPATH% -cp %NDCLASSPATH% com.ibm.internet.nd.framework.FWK_Framework com.ibm.internet.nd.gui.GUI_eNDRootNode1
$JREDIR/$JRE -mx64m $END_ACCESS $CONFIG_DIR -DEND_INSTALL_PATH=/opt/&BASEDIR -cp $NDCLASSPATH com.ibm.internet.nd.framework.FWK_Framework com.ibm.internet.nd.gui.GUI_eNDRootNode &1
re -mx64m $END_ACCESS $CONFIG_DIR $NDLOCALE -DEND_INSTALL_PATH=/opt/nd -classpath $NDCLASSPATH com.ibm.internet.nd.framework.FWK_Framework com.ibm.internet.nd.gui.GUI_eNDRootNode 1>/dev/null 2>&1 &1
ava -mx64m $END_ACCESS $CONFIG_DIR $NDLOCALE -DEND_INSTALL_PATH=/usr/lpp/&BASEDIR -classpath $NDCLASSPATH com.ibm.internet.nd.framework.FWK_Framework com.ibm.internet.nd.gui.GUI_eNDRootNode 1>/dev/null 2>&1 &
There is no recommended value for n , but it should be greater than he default option. A good place to start would be with twice the default value.
This problem can occur when another application is using one of the ports used by CBR. For more information, go to Checking CBR port numbers.
The cbrcontrol command returns: "Error: Server not responding." Or, the ndadmin command returns: "Error: unable to access RMI server." These errors can result when your machine has a socksified stack. To correct this problem, edit the socks.cnf file to contain the following lines:
EXCLUDE-MODULE java EXCLUDE-MODULE jre EXCLUDE-MODULE jrew EXCLUDE-MODULE javaw
These errors can also occur if you have not already started cbrserver.
Caching Proxy and CBR have been started, but requests are not being load balanced. This error can occur if you start Caching Proxy before starting the executor. If this happens, the stderr log for Caching Proxy will contain the following error message: "ndServerInit: Could not attach to executor." To avoid this problem, start the executor before starting Caching Proxy.
On Solaris, the cbrcontrol executor start command returns: "Error: Executor was not started." This error occurs if you do not configure the IPC (Inter-process Communication) for the system so that the maximum size of a shared memory segment and semaphore IDs are bigger than the operating system's default. In order to increase the size of the shared memory segment and semaphore IDs, you must edit the /etc/system file. For more information on how to configure this file, see ***.
If the URL rule does not work, this can be a result of either a syntactical or configuration error. For this problem check the following:
This problem can occur when another application is using one of the ports used by Mailbox Locator. For more information, go to Checking Mailbox Locator port numbers.
On a UNIX platform, this problem occurs when mlserver is used to load balance a large number of IMAP/POP3 client requests and the system's limit on file descriptors is too small for the number of requests that mlserver is trying to service. The mlserver produces the following exception and then is stopped:
java.rmi.RMISecurityException: security.fd.read
The protocol specific proxy log file reports:
SocketException=java.net.SocketException: Socket closed
The solution is to modify the nofiles (AIX, Linux) or the open files (Solaris) limit in the shell where mlserver is started. Increase the nofiles limit to a reasonable number larger than the current nofiles limit. Use ulimit -a to display the current nofiles limit, and use ulimit -n value to increase the value.
The mlcontrol command returns: "Error: Server not responding." Or, the ndadmin command returns: "Error: unable to access RMI server." These errors can result when your machine has a socksified stack. To correct this problem, edit the socks.cnf file to contain the following lines:
EXCLUDE-MODULE java EXCLUDE-MODULE jre EXCLUDE-MODULE jrew EXCLUDE-MODULE javaw
These errors can also occur if you have not already started mlserver.
When attempting to add a port to a configuration, you might receive this error message: Error: Unable to add port. It is possible that another application is already listening on that port. Mailbox Locator tries to start a proxy that binds to the cluster IP on the specified port in the command. If another application is binding to that IP or listening to all IPs on that port, the proxy startup fails. To use Mailbox Locator on that port, you must stop the conflicting application.
Note that on the Linux platform, the xinetd daemon can start a listener without running, for example, a POP3 program. It is, therefore, important to check netstat -a to determine if any application is listening on the intended port.
For Mailbox Locator, the mlcontrol port add command produces the following error message: "The proxy on cluster <cluster>, port <port> did not start." The solution is to configure the cluster address on a NIC before the proxy can be started. Also, verify that no other application is running on that port listening for the clustor address (including a general listen-on-everything application).
This problem can occur when another application is using one of the ports used by Site Selector. For more information, go to Checking Site Selector port numbers.
Symptom: Site Selector component does not round-robin incoming requests from Solaris clients.
Possible cause: Solaris systems run a name service cache daemon. If this daemon is running, the subsequent resolver request will be answered from this cache instead of querying Site Selector.
Solution: Turn off the name service cache daemon on the Solaris machine.
The sscontrol command returns: "Error: Server not responding." Or, the ndadmin command returns: "Error: unable to access RMI server." These errors can result when your machine has a socksified stack. To correct this problem, edit the socks.cnf file to contain the following lines:
EXCLUDE-MODULE java EXCLUDE-MODULE jre EXCLUDE-MODULE jrew EXCLUDE-MODULE javaw
These errors can also occur if you have not already started ssserver.
Site Selector must be able to participate in a DNS. All the machines involved in the configuration should also be participants of this system. Windows does not always require the configured host name to be in the DNS. Site Selector requires that its host name be define in the DNS to start properly.
Verify this host is defined in the DNS. Edit the ssserver.cmd file and remove the "w" from "javaw". This should provide more errors.
Site Selector's name server does not bind to any one address on the machine. It will respond to requests destined for any valid IP on the machine. Site Selector relies on the operating system to route the response back to the client. If the Site Selector machine has multiple adapters and any number of them are attached to the same subnet, it is possible the O/S will send the response to the client from a different address than it was received. Some client applications will not accept a response received from an address other than where it was sent. As a result, the name resolution will appear to fail.
This problem can occur when another application is using one of the ports used by the Consultant's lbcserver. For more information, see Checking Cisco Consultant port numbers.
The lbccontrol command returns: "Error: Server not responding." Or, the ndadmin command returns: "Error: unable to access RMI server." These errors can result when your machine has a socksified stack. To correct this problem, edit the socks.cnf file to contain the following lines:
EXCLUDE-MODULE java EXCLUDE-MODULE jre EXCLUDE-MODULE jrew EXCLUDE-MODULE javaw
These errors can also occur if you have not already started lbcserver.
This problem can occur when a valid product license is missing. When you attempt to start lbcserver, you receive the following message:
Your license has expired. Contact your local IBM representative or authorized IBM reseller.
To correct this problem:
You must use the full metric name for user-written metrics on Windows 2000 Metric Servers. For example, you must specify usermetric.bat instead of usermetric. The name usermetric is valid on the command line, but will not work when executed from within the runtime environment. If you do not use the full metric name, you will receive a Metric Server IOException. Set the LOG_LEVEL variable to a value of 3 in the metricserver command file, then check the log output. In this example, the exception appears as:
... java.io.IOException: CreateProcess: usermetric error=2
There can be several reasons why Metric Server is not reporting load information to Network Dispatcher. To determine the cause, perform the following checks:
The Metric Server log reports this error message after key files have been transferred to the server.
This error is logged when the key file fails authorization with the paired key due to corruption in the pair. To correct this problem try the following:
The syntax diagram shows you how to specify a command so that the operating system can correctly interpret what you type. Read the syntax diagram from left to right and from top to bottom, following the horizontal line (the main path).
The following symbols are used in syntax diagrams:
You must include all punctuation such as colons, quotation marks, and minus signs that are shown in the syntax diagram.
The following types of parameters are used in syntax diagrams.
Parameters are classified as keywords or variables. Keywords are displayed in lowercase letters and can be entered in lowercase. For example, a command name is a keyword. Variables are italicized and represent names or values you supply.
In the following example, the user command is a keyword. The required variable is user_id, and the optional variable is password. Replace the variables with your own values.
>>-user--user_id--+----------+--------------------------------->< '-password-'
Required keywords: required keywords and variables appear on the main path line.
>>-required_keyword--------------------------------------------><
You must code required keywords and values.
Choose one required item from a stack: If there is more than one mutually exclusive required keyword or variable to choose from, they are stacked vertically in alphanumeric order.
>>-+-required_parameter_1-+------------------------------------>< '-required_parameter_2-'
Optional values: Optional keywords and variables appear below the main path line.
>>-+---------+------------------------------------------------->< '-keyword-'
You can choose not to code optional keywords and variables.
Choose one optional keyword from a stack: If there is more than one mutually exclusive optional keyword or variable to choose from, they are stacked vertically in alphanumeric order below the main path line.
>>-+-------------+--------------------------------------------->< +-parameter_1-+ '-parameter_2-'
Variables: A word in all italics is a variable. Where you see a variable in the syntax, you must replace it with one of its allowable names or values, as defined in the text.
>>-variable----------------------------------------------------><
Nonalphanumeric characters: If a diagram shows a character that is not alphanumeric (such as colons, quotes, or minus signs), you must code the character as part of the syntax. In this example, you must code cluster:port.
>>-cluster:port------------------------------------------------><
This appendix describes how to use the Dispatcher ndcontrol commands. It is also a command reference for CBR and Mailbox Locator. CBR and Mailbox Locator use a subset of the Dispatcher commands. See Configuration differences between CBR, Mailbox Locator, and Dispatcher for more information.
Below is a list of commands in this appendix:
You can enter a minimized version of the ndcontrol command parameters. You only need to enter the unique letters of the parameters. For example, to get help on the file save command, you can type ndcontrol he f instead of ndcontrol help file.
To start up the command line interface: issue ndcontrol to receive an ndcontrol command prompt.
To end the command line interface: issue exit or quit.
The CBR and Mailbox Locator command line interface is for the most part a subset of the command line interface of Dispatcher. Use the cbrcontrol command (for the CBR component) or use the mlcontrol command (for the Mailbox Locator component) instead of ndcontrol to configure the component.
Some of the commands that are omitted in CBR are listed below.
Some of the commands that are omitted in Mailbox Locator are listed below.
>>-ndcontrol--advisor--+-connecttimeout--name--+-port-----+---timeoutseconds--+-> | '-cluster:port-' | +-interval--name--+-port-----+---seconds---------------+ | '-cluster:port-' | +-list-------------------------------------------------+ +-loglevel--name--+-port-----+---level-----------------+ | '-cluster:port-' | +-logsize--name--+-port-----+---+-unlimited---------+--+ | '-cluster:port-' '-number of records-' | +-receivetimeout--name--+-port-----+---timeoutseconds--+ | '-cluster:port-' | +-report--name--+-port-----+---------------------------+ | '-cluster:port-' | +-start--name--+-port-----+---+-----------+------------+ | '-cluster:port-' '-log file--' | +-status--name--+-port-----+---------------------------+ | '-cluster:port-' | +-stop--name--+-port-----+-----------------------------+ | '-cluster:port-' | +-timeout--name--+-port-----+---+-unlimited-+----------+ | '-cluster:port-' '-seconds---' | '-version--name--+-port-----+--------------------------' '-cluster:port-' >--------------------------------------------------------------><
Names of customized advisors are of the format xxxx, where ADV_xxxx is the name of the class that implements the custom advisor. See Create custom (customizable) advisors for more information.
The cluster is the address in dotted-decimal format or symbolic name. The port is the number of the port that the advisor is monitoring.
Advisor Name | Protocol | Port |
---|---|---|
connect | ICMP | 12345 |
db2 | private | 50000 |
dns | DNS | 53 |
ftp | FTP | 21 |
http | HTTP | 80 |
ibmproxy | HTTP (via Caching Proxy) | 80 |
imap | IMAP | 143 |
nntp | NNTP | 119 |
ping | PING | 0 |
pop3 | POP3 | 110 |
self | private | 12345 |
smtp | SMTP | 25 |
ssl | HTTP | 443 |
ssl2http | SSL | 443 |
telnet | Telnet | 23 |
WLM | private | 10,007 |
The default file is advisorname_port.log, for example, http_80.log. To change the directory where the log files will be kept, see Changing the log file paths. The default log files for cluster (or site) specific advisors are created with the cluster address, for example, http_127.40.50.1_80.log.
Examples
ndcontrol advisor start http 127.40.50.1:80
ndcontrol advisor start http 88
ndcontrol advisor stop http 127.40.50.1:80
ndcontrol advisor connecttimeout http 80 30
ndcontrol advisor connecttimeout http 127.40.50.1:80 20
ndcontrol advisor interval ftp 21 6
ndcontrol advisor listThis command produces output similar to:
--------------------------------------- | ADVISOR | CLUSTER:PORT | TIMEOUT | --------------------------------------- | http |127.40.50.1:80 | unlimited | | ftp | 21 | unlimited | ---------------------------------------
ndcontrol advisor loglevel http 80 0
ndcontrol advisor logsize ftp 21 5000
ndcontrol advisor receivetimeout http 80 60
ndcontrol advisor report ftp 21This command produces output similar to:
Advisor Report: --------------- Advisor name ............. Ftp Port number .............. 21 Cluster address .......... 9.67.131.18 Server address ........... 9.67.129.230 Load ..................... 8 Cluster address .......... 9.67.131.18 Server address ........... 9.67.131.215 Load ..................... -1
ndcontrol advisor status http 80This command produces output similar to the following:
Advisor Status: --------------- Interval (seconds) ............ 7 Timeout (seconds) ............. Unlimited Connect timeout (seconds).......21 Receive timeout (seconds).......21 Advisor log filename .......... Http_80.log Log level ..................... 1 Maximum log size (bytes) ...... Unlimited
ndcontrol advisor timeout ftp 21 5
ndcontrol advisor version ssl 443This command produces output similar to the following:
Version: 04.00.00.00 - 07/12/2001-10:09:56-EDT
>>-ndcontrol--cluster--+-add--cluster+c2+...--+-----------------------------------------+-+-> | +-proportions--active--new--port--system--+ | | +-maxports--size--------------------------+ | | +-maxservers--size------------------------+ | | +-stickytime--time------------------------+ | | +-weightbound--weight---------------------+ | | +-porttype--type--------------------------+ | | +-primaryhost--address--------------------+ | | +-staletimeout--staletimeout--------------+ | | '-sharedbandwidth--size-------------------' | +-set--cluster+c2+...--+-maxports--size--------------+-------------+ | +-maxservers--size------------+ | | +-stickytime--time------------+ | | +-weightbound--weight---------+ | | +-porttype--type--------------+ | | +-primaryhost--address--------+ | | +-staletimeout--staletimeout--+ | | '-sharedbandwidth--size-------' | +-remove--cluster--------------------------------------------------+ +-report--cluster--------------------------------------------------+ +-status--cluster--------------------------------------------------+ +-configure--cluster--+-------------------------+------------------+ | '-interfacename - netmask-' | '-unconfigure--cluster---------------------------------------------' >--------------------------------------------------------------><
With the exception of the ndcontrol cluster add command, you can use a colon (:) to act as a wild card. For example, the following command, ndcontrol cluster set : weightbound 80, will result in setting a weightbound of 80 to all clusters.
If you change the primaryhost of a cluster once the primary and backups are already started and running mutual high availability, you also must force the new primary host to takeover. And, you need to update the scripts and manually unconfigure and configure the cluster correctly. See Mutual high availability for more information.
Examples
ndcontrol cluster add 130.40.52.153
ndcontrol cluster remove 130.40.52.153
ndcontrol cluster set 9.6.54.12 proportions 60 35 5 0
ndcontrol cluster add 0.0.0.0
ndcontrol cluster set 9.6.54.12 primaryhost 9.65.70.19
ndcontrol cluster status 9.67.131.167This command produces output similar to:
Cluster Status: --------------- Address ................................. 9.67.131.167 Number of target ports .................. 3 Default sticky time ..................... 0 Default stale timeout ................... 30 Default port weight bound ............... 20 Maximum number of ports ................. 8 Default port protocol ................... tcp/udp Default maximum number of servers ....... 32 Proportion given to active connections... 0.5 Proportion given to new connections...... 0.5 Proportion given specific to the port.... 0 Proportion given to system metrics....... 0 Shared bandwidth (KBytes) ............... 0 Primary Host Address .................... 9.67.131.167
>>-ndcontrol--executor--+-report-------------------------------+-> +-set--+-nfa--IP address-------------+-+ | +-maxclusters--size-----------+ | | +-maxports--size--------------+ | | +-fincount--fincount----------+ | | +-fintimeout--fintimeout------+ | | +-maxservers--size------------+ | | +-staletimeout--staletimeout--+ | | +-stickytime--time------------+ | | +-clientgateway--address------+ | | +-weightbound--weight---------+ | | +-porttype--type--------------+ | | +-wideportnumber--port--------+ | | '-sharedbandwidth--size-------' | +-start--------------------------------+ +-status-------------------------------+ '-stop---------------------------------' >--------------------------------------------------------------><
Examples
ndcontrol executor status Executor Status: ---------------- Nonforwarding address ............... 9.67.131.151 Client gateway address .............. 0.0.0.0 Fin count ........................... 4,000 Fin timeout ......................... 60 Wide area network port number ....... 2,001 Shared bandwidth (Kbytes) ........... 0 Default maximum ports per cluster ... 8 Maximum number of clusters .......... 100 Default maximum servers per port .... 32 Port stale timeout .................. 300 Port sticky time .................... 0 Port weight bound ................... 20 Maximum number of clusters .......... 100
ndcontrol executor set nfa 130.40.52.167
ndcontrol executor set maxclusters 4096
ndcontrol executor start
ndcontrol executor stop
>>-ndcontrol--file--+-delete--file[.ext]-----------+----------->< +-appendload--file[.ext]-------+ +-report-----------------------+ +-save--file[.ext]--+--------+-+ | '-force--' | '-newload--file[.ext]----------'
The file extension (.ext) can be anything you like and can be omitted.
Common install directory path -- c:\Program Files\ibm\edge\nd\servers\configurations\component
Native install directory path -- c:\Program Files\ibm\nd\servers\configurations\component
Examples
ndcontrol file delete file3 File (file3) was deleted.
ndcontrol file newload file1.sv File (file1.sv) was loaded into the Dispatcher.
ndcontrol file appendload file2.sv File (file2.sv) was appended to the current configuration and loaded.
ndcontrol file report FILE REPORT: file1.save file2.sv file3
ndcontrol file save file3 The configuration was saved into file (file3).
>>-ndcontrol--help--+-help--------------+---------------------->< +-host--------------+ +-executor----------+ +-manager-----------+ +-advisor-----------+ +-cluster-----------+ +-port--------------+ +-rule--------------+ +-server------------+ +-subagent----------+ +-highavailability--+ +-file--------------+ +-set---------------+ +-status------------+ '-log---------------'
Examples
ndcontrol helpThis command produces output similar to:
HELP COMMAND ARGUMENTS: --------------------------------- Usage: help <help option> Example: help cluster help - print complete help text advisor - help on advisor command cluster - help on cluster command executor - help on executor command file - help on file command host - help on host command log - help on log command manager - help on manager command metric - help on metric command port - help on port command rule - help on rule command server - help on server command set - help on set command status - help on status command subagent - help on subagent command highavailability - help on high availability commandNotice that parameters within <> are variables.
fintimeout <cluster address>|all <time> -Change FIN timeout (Use 'all' to change all clusters)
>>-ndcontrol--highavailability--+-status----------------------------------------+-> +-backup--+-add--+-primary-+---+-auto---+---p-+-+ | | +-backup--+ '-manual-' | | | | '-both----' | | | '-delete----------------------------' | +-reach--+-add----+---address--mask-------------+ | '-delete-' | +-heartbeat--+-add--srcaddress--dstaddress--+---+ | '-delete--address--------------' | '-takeover--+---------+-------------------------' '-address-' >--------------------------------------------------------------><
In addition, the status keyword returns information about various substates:
Mutual high availability configuration (role of each Dispatcher machine is both):
Notes:
Examples
ndcontrol highavailability status
Output:
High Availability Status: ------------------------- Role ........................primary Recovery Strategy ........... manual State ....................... Active Sub-state.............. Synchronized Primary host........... 9.67.131.151 Port .........................12,345 Preferred Target....... 9.67.134.223 Heartbeat Status: ----------------- Count ................ 1 Reachability Status: -------------------- Count ................ 1
ndcontrol highavailability backup add primary auto 80
ndcontrol highavailability reach add 9.67.125.18
Primary - highavailability heartbeat add 9.67.111.3 9.67.186.8 Backup - highavailability heartbeat add 9.67.186.8 9.67.111.3
ndcontrol highavailability takeover
>>-ndcontrol--host:--remote_host-------------------------------><
ndcontrol host:remote_host
After this command has been issued on the command prompt, enter any valid ndcontrol command you want issued to the remote Network Dispatcher machine.
>>-ndcontrol--log--+-start-------------------------+----------->< +-stop--------------------------+ +-set--+-retention-hours------+-+ | '---interval-seconds---' | '-status------------------------'
>>-ndcontrol--manager--+-interval--seconds-------------------+->< +-loglevel--level---------------------+ +-logsize--+-unlimited-+--------------+ | '-bytes-----' | +-quiesce--server--+------+-----------+ | '-now--' | +-reach set--+-interval-seconds-----+-+ | '-+-loglevel-level---+-' | | '---logsize-size---' | +-refresh--refresh cycle--------------+ +-report--+-----------------+---------+ | '-cluster+c2+...--' | +-restart--message--------------------+ +-sensitivity--weight-----------------+ +-smoothing--smoothing index----------+ +-start--+------------------------+---+ | '-log file--metric_port--' | +-status------------------------------+ +-stop--------------------------------+ +-unquiesce--server-------------------+ '-version-----------------------------'
Or, if you used server partitioning, use the logical server's unique name. See Server Partitioning: logical servers configured to one physical server (IP address) for more information.
The default file will be installed in the logs directory. See Appendix F, Sample configuration files. To change the directory where the log files will be kept, see Changing the log file paths.
Examples
ndcontrol manager interval 5
ndcontrol manager loglevel 0
ndcontrol manager logsize 1000000
ndcontrol manager quiesce 130.40.52.153
ndcontrol manager refresh 3
ndcontrol manager reportThis command produces output similar to:
---------------------------------- | HOST TABLE LIST | STATUS | ---------------------------------- | 9.67.129.221| ACTIVE| | 9.67.129.213| ACTIVE| | 9.67.134.223| ACTIVE| ---------------------------------- ----------------------------------------------------------------------------------- | 9.67.131.18| WEIGHT | ACTIVE % 48| NEW % 48 | PORT % 4 | SYSTEM % 0 | ----------------------------------------------------------------------------------- | PORT: 80 |NOW | NEW |WT | CONN |WT | CONN | WT | LOAD | WT | LOAD | ----------------------------------------------------------------------------------- | 9.67.129.221| 8 | 8 | 10 | 0| 10 | 0 | 7 | 29 | 0 | 0| | 9.67.134.223| 11 | 11 | 10 | 0| 10 | 0 | 12 | 17 | 0 | 0| ----------------------------------------------------------------------------------- | PORT TOTALS:| 19 | 19 | | 0| | 0 | | 46 | | 0| ----------------------------------------------------------------------------------- ----------------------------------------------------------------------------------- | 9.67.131.18| WEIGHT | ACTIVE % 48| NEW % 48 | PORT % 4 | SYSTEM % 0 | ----------------------------------------------------------------------------------- | PORT: 23 |NOW | NEW |WT | CONN |WT | CONN | WT | LOAD | WT | LOAD | ----------------------------------------------------------------------------------- | 9.67.129.213| 10 | 10 | 10 | 0| 10 | 0 | 10 | 71 | 0 | 0| | 9.67.134.223| 0 | 0 | 10 | 0| 10 | 0 |-9999 | -1 | 0 | 0| ----------------------------------------------------------------------------------- | PORT TOTALS:| 10 | 10 | | 0| | 0 | | 70 | | 0| ----------------------------------------------------------------------------------- -------------------------------- | ADVISOR | PORT | TIMEOUT | -------------------------------- | reach | 0 | unlimited | | http | 80 | unlimited | | ftp | 21 | unlimited | --------------------------------
ndcontrol manager restart Restarting the manager to update codeThis command produces output similar to:
320-14:04:54 Restarting the manager to update code
ndcontrol manager sensitivity 10
ndcontrol manager smoothing 2.0
ndcontrol manager start ndmgr.log
ndcontrol manager statusThis command produces output similar to the following example.
Manager status: ============= Metric port................................... 10,004 Manager log filename.......................... manager.log Manager log level............................. 1 Maximum manager log size (bytes).............. unlimited Sensitivity level............................. 0.05 Smoothing index............................... 1.5 Update interval (seconds)..................... 2 Weights refresh cycle......................... 2 Reach log level............................... 1 Maximum reach log size (bytes)................ unlimited Reach update interval (seconds)............... 7
ndcontrol manager stop
ndcontrol manager quiesce 130.40.52.153 now
ndcontrol manager quiesce 130.40.52.153
ndcontrol manager unquiesce 130.40.52.153
ndcontrol manager version
>>-ndcontrol--metric--+-add--cluster+c2+...+cN:metric+metric1+...+metricN---------------+-> +-remove--cluster+c2+...+cN:metric+metric1+...+metricN------------+ +-proportions--cluster+c2+...+cN proportion1 prop2 prop3...propN--+ '-status--cluster+c2+...+cN:metric+metric1+...+metricN------------' >--------------------------------------------------------------><
Note: For Cisco Consultant, the cluster address corresponds to the virtual IP (VIP) address of the content rule of the owner in the Cisco CSS Switch configuration.
Examples
sscontrol metric add site1:metric1
sscontrol metric proportions site1 0 100
sscontrol metric status site1:metric1This command produces output similar to the following:
Metric Status: ------------ Cluster ....................... 10.10.10.20 Metric name ................... metric1 Metric proportion ............. 50 Server .................... plm3 Metric data ............... -1
>>-ndcontrol--port--+-add--cluster:port--+-----------------------+-+-> | +-crossport--otherport--+ | | +-maxservers--size------+ | | +-stickymask--value-----+ | | +-stickytime--time------+ | | +-method--type----------+ | | +-staletimeout--value---+ | | +-weightbound--weight---+ | | +-porttype--type--------+ | | '-protocol--type--------' | +-set--cluster:port--+-crossport--otherport--+-+ | +-maxservers--size------+ | | +-stickymask--value-----+ | | +-stickytime--time------+ | | +-staletimeout--value---+ | | +-weightbound--weight---+ | | +-porttype--type--------+ | | '-maxhalfopen--value----' | +-remove--cluster:port-------------------------+ +-report--cluster:port-------------------------+ +-status--cluster:port-------------------------+ '-halfopenaddressreport--cluster:port----------' >--------------------------------------------------------------><
On Windows, this means it must be in the Windows Networking setup. The cluster configure command is not enough, because it only simulates IP aliasing, and the proxy cannot bind to this fake IP. For all other operating systems, the cluster configure command is appropriate because it uses ifconfig to alias the IP.
To remove the crossport feature, set the crossport value back to its own port number. For more information on cross port affinity feature, see Cross port affinity.
For the Dispatcher component:
For the CBR component: If you set the port stickytime to a nonzero value, then the affinity type on the rule must be none (default). Rule-based affinity (passive cookie, URI, active cookie) cannot co-exist when stickytime is set on the port.
A positive value indicates that a check will be made to determine if the current half-open connections exceeds the threshold. If the current value is above the threshold, a call to an alert script is made. See Denial of service attack detection for more information.
Examples
ndcontrol port add 130.40.52.153:80+23
ndcontrol port set 130.40.52.153:0
mlcontrol port add 9.37.60.91:20 protocol pop3
ndcontrol port set 130.40.52.153:80 weightbound 10
ndcontrol port set 130.40.52.153:80+23 stickytime 60
ndcontrol port set 130.40.52.153:80 crossport 23
ndcontrol port remove 130.40.52.153:23
ndcontrol port status 9.67.131.153:80This command produces output similar to:
Port Status: ------------ Port number .................... 80 Cluster address ................ 9.67.131.153 Number of servers .............. 2 Stale timeout .................. 30 Weight bound ................... 20 Maximum number of servers ...... 32 Sticky time .................... 0 Port type ...................... tcp/udp Forwarding method .............. MAC Based Forwarding Sticky mask bits ............... 32 Cross Port Affinity ............ 80 Max Half Open Connections ...... 0
ndcontrol port halfopenaddressreport 9.67.127.121:80This command produces output similar to:
Half open connection report successfully created: ------------ Half Open Address Report for cluster:port = 9.67.127.121:80 Total addresses with half open connections reported ... 0 Total number of half open connections reported ........ 0 Largest number of half open connections reported ...... 0 Average number of half open connections reported ...... 0 Average half open connection time (seconds) reported .. 0 Total half open connections received .................. 0
>>-ndcontrol--rule--+-add--cluster:port:rule--type--type--| opts |--+-> +-dropserver--cluster:port:rule--server---------+ +-remove--cluster:port:rule---------------------+ +-report--cluster:port:rule---------------------+ +-set--cluster:port:rule--| opts |--------------+ +-status--cluster:port:rule---------------------+ '-useserver--cluster:port:rule--server+s2+...---' >-------------------------------------------------------------->< opts |---+----------------------------------+------------------------| +-beginrange--low--endrange--high--+ +-priority--level------------------+ +-pattern=--pattern----------------+ +-tos--value-----------------------+ +-stickytime--time-----------------+ +-affinity--affinity_type----------+ +-cookiename--value----------------+ +-evaluate--level------------------+ '-sharelevel--level----------------'
See Active cookie affinity for more information.
An affinity type of "activecookie" enables load-balancing Web traffic with affinity to the same server based upon cookies generated by Network Dispatcher.
An affinity type of "passivecookie" enables load-balancing Web traffic with affinity to the same server based upon self-identifying cookies generated by the servers. You must use the cookiename parameter in conjunction with passive cookie affinity.
An affinity type of "URI" enables load-balancing Web traffic to caching-proxy servers in a manner which effectively increases the size of the cache.
See Active cookie affinity, Passive cookie affinity, and URI affinity for more information.
See Passive cookie affinity for more information.
Or, if you used server partitioning, use the logical server's unique name. See Server Partitioning: logical servers configured to one physical server (IP address) for more information.
Examples
ndcontrol rule add 9.37.67.100:80:trule type true priority 100
ndcontrol rule add 9.37.131.153:80:ni type ip b 9.0.0.0 e 9.255.255.255
ndcontrol rule add cluster1:80:timerule type time beginrange 11 endrange 14 ndcontrol rule useserver cluster1:80:timerule server05
ndcontrol rule add 9.67.131.153:80:tosrule type service tos 0xx1001x
ndcontrol rule add 9.67.131.153:80:rbwrule type reservedbandwidth beginrange 0 endrange 100 evaluate rule
ndcontrol cluster set 9.67.131.153 sharedbandwidth 200 ndcontrol rule add 9.67.131.153:80:shbwrule type sharedbandwidth sharelevel cluster
>>-ndcontrol--server--+-add--cluster:port:server--+--------------------------+-+-> | +-address--address---------+ | | +-collocated--value--------+ | | +-sticky--value------------+ | | +-weight--value------------+ | | +-fixedweight--value-------+ | | +-mapport--portvalue-------+ | | +-router--addr-------------+ | | +-cookievalue--value-------+ | | +-returnaddress--addr------+ | | +-advisorrequest--string---+ | | '-advisorresponse--string--' | +-set--cluster:port:server--+-collocated--value--------+-+ | +-sticky--value------------+ | | +-weight--value------------+ | | +-fixedweight--value-------+ | | +-router--addr-------------+ | | +-cookievalue--value-------+ | | +-advisorrequest--string---+ | | '-advisorresponse--string--' | +-down--cluster:port:server------------------------------+ +-remove--cluster:port:server----------------------------+ +-report--cluster:port:server----------------------------+ +-up--cluster:port:server--------------------------------+ '-status--cluster:port:server----------------------------' >--------------------------------------------------------------><
Or, if you use a unique name that does not resolve to an IP address, you must provide the server address parameter on the ndcontrol server add command. See Server Partitioning: logical servers configured to one physical server (IP address) for more information.
Examples
ndcontrol server add 130.40.52.153:80:27.65.89.42
ndcontrol server set 130.40.52.153:80:27.65.89.42 sticky no
ndcontrol server down 130.40.52.153:80:27.65.89.42
ndcontrol server remove ::27.65.89.42
ndcontrol server set 130.40.52.153:80:27.65.89.42 collocated yes
ndcontrol server set 130.40.52.153:80:27.65.89.42 weight 10
ndcontrol server up 130.40.52.153:80:27.65.89.42
ndcontrol server add 130.40.52.153:80:130.60.70.1 router 130.140.150.0
ndcontrol server set 130.40.52.153:80:27.65.89.42 advisorrequest "\"HEAD / HTTP/2.0\""
>>-ndcontrol--set--+-loglevel--level--------+------------------>< +-logsize--+-unlimited-+-+ | '-size------' | '-logstatus--------------'
>>-ndcontrol--status-------------------------------------------><
Examples
ndcontrol statusThis command produces output similar to:
Executor has been started. Manager has been started. -------------------------------- | ADVISOR | PORT | TIMEOUT | -------------------------------- | reach | 0 | unlimited | | http | 80 | unlimited | | ftp | 21 | unlimited | --------------------------------
>>-ndcontrol--subagent--+-loglevel--level---------------------+->< +-logsize--+-bytes-----+--------------+ | '-unlimited-' | +-report------------------------------+ +-start--+--------------------------+-+ | '-community_name--logfile--' | +-status------------------------------+ +-stop--------------------------------+ '-version-----------------------------'
Examples
ndcontrol subagent start bigguy bigguy.log
This appendix describes how to use the content rule (pattern) syntax for the CBR component and the Dispatcher component's cbr forwarding method, along with scenarios and examples of their usage.
Only applicable if you selected "content" for the rule type.
Enter the pattern syntax you want to use, using the following restrictions
Reserved keywords are always followed by an equal sign "=".
A browser targeting http://www.company.com/path/webpage.htm might result in values such as:
Method=GET URI=/path/webpage.htm Version=/HTTP/1.1 Host=www.company.com Connection=Keep-Alive Referer=http://www.company.com/path/parentwebpage.htm
For example, the following command is valid only when using the cbrcontrol>> prompt.
rule add 10.1.203.4:80:cbr_prod_rule_ek type content pattern client=181.0.153.222&uri=http://10.1.203.4/nipoek/*
When using special characters, for this same command to work at the operating system's prompt, double quotation marks (" ") must be placed around the pattern as follows:
cbrcontrol rule add 10.1.203.4:80:cbr_prod_rule_ek type content pattern "client=181.0.153.222&uri=http://10.1.203.4/nipoek/*"
If the quotation marks are not used, some of the pattern might be truncated when the rule is saved in CBR. Note that quotation marks are not supported when using the cbrcontrol>> command prompt.
The following is a collection of possible scenarios and examples for using pattern syntaxes
Scenario 1:
The setup for one cluster name involves one set of Web servers for standard HTML content, another set of Web servers with WebSphere Application Server for servlet requests, another set of Lotus Notes servers for NSF files, and so forth. Access to the client data is required to distinguish between those requested pages. It is also required to send them to the appropriate servers. The content pattern matching rules provide the separation needed to accomplish these tasks. A series of rules are configured so that the necessary separation of requests occurs automatically. For example, the following commands accomplish the three splits mentioned:
>>rule add cluster1:80:servlets type content pattern uri=*/servlet/* priority 1 >>rule uses cluster1:80:servlets server1+server2
>>rule add cluster1:80:notes type content pattern uri=*.nsf* priority 2 >>rule uses cluster1:80:notes server3+server4
>>rule add cluster1:80:regular type true priority 3 >>rule uses cluster1:80:regular server5+server6
If a request for an NSF file arrives at Network Dispatcher, the servlets rule is checked first, but does not match. The request is then checked by the notes rule and returns a match. The client is load-balanced between server3 and server4.
Scenario 2
Another common scenario is when the main Web site controls several distinct internal groups. For example, www.company.com/software involves a different set of servers and content from www.company.com/hardware division. Because the requests are all based off the root www.company.com cluster, content rules are required to find the URI differences and complete load balancing. The scenario's rule looks similar to the following:
>>rule add cluster1:80:div1 type content pattern uri=/software/* priority 1 >>rule uses cluster1:80:div1 server1+server2
>>rule add cluster1:80:div2 type content pattern uri=/hardware/* priority 2 >>rule uses cluster1:80:div2 server3+server4
Scenario 3
Certain combinations are sensitive to the order in which rules are searched. For example, in Scenario 2, clients were split based on a directory in their request path; however, the target directory might appear at multiple levels of the path and mean different things on placement. For example, www.company.com/pcs/fixes/software is a different target from www.company.com/mainframe/fixes/software. The rules must be defined to account for this possibility and not catch too many scenarios at the same time. For example, the "uri=*/software/*" test is too broad a wildcard search in this case. Alternative rules could be structured in the following manner:
A combination search can narrow this down:
>>rule add cluster1:80:pcs type content pattern (uri=/pcs/*)&(uri=*/software/*) >>rule uses cluster 1:80:pcs server1
In cases where there are no combinations to use, the order becomes important:
>>rule add cluster1:80:pc1 type content pattern uri=/pcs/* >>rule uses cluster1:80:pc1 server2
The second rule catches when "pcs" appears in later directory spots instead of the first.
>>rule add cluster1:80:pc2 type content pattern uri=/*/pcs/* >>rule uses cluster1:80:pc2 server3
In almost every case, you want to complete the rules with a default always true rule to catch anything that falls through the other rules. This can also be a "Sorry, the site is currently down, please try again later" server for scenarios where all other servers fail for this client.
>>rule add cluster1:80:sorry type true priority 100 >>rule uses cluster1:80:sorry server5
This appendix describes how to use the following Site Selector sscontrol commands:
You can enter a minimized version of the sscontrol command parameters. You only need to enter the unique letters of the parameters. For example, to get help on the file save command, you can enter sscontrol he f instead of sscontrol help file.
>>-sscontrol--advisor--+-connecttimeout--name--+-port------+---seconds-------+-> | '-sitename:port-' | +-interval--name--+-port------+---seconds-------------+ | '-sitename:port-' | +-list------------------------------------------------+ +-loglevel--name--+-port------+---level---------------+ | '-sitename:port-' | +-logsize--name--+-port------+---+-size | unlimited-+-+ | '-sitename:port-' '-bytes------------' | +-receivetimeout--name--+-port------+---seconds-------+ | '-sitename:port-' | +-report--name--+-port------+-------------------------+ | '-sitename:port-' | +-start--name--+-port------+---+-----------+----------+ | '-sitename:port-' '-log file--' | +-status--name--+-port------+-------------------------+ | '-sitename:port-' | +-stop--name--+-port------+---------------------------+ | '-sitename:port-' | +-timeout--name--+-port------+------------------------+ | '-sitename:port-' | '-version--name--+-port------+---seconds--------------' '-sitename:port-' >--------------------------------------------------------------><
.
Advisor Name | Protocol | Port |
---|---|---|
Connect | n/a | user-defined |
db2 | private | 50000 |
ftp | FTP | 21 |
http | HTTP | 80 |
imap | IMAP | 143 |
nntp | NNTP | 119 |
PING | PING | 0 |
pop3 | POP3 | 110 |
smtp | SMTP | 25 |
ssl | SSL | 443 |
telnet | Telnet | 23 |
The default file is advisorname_port.log, for example, http_80.log. To change the directory where the log files are stored, see Changing the log file paths.
You can start only one advisor for each sitename.
Examples
sscontrol advisor connecttimeout http 80 30
sscontrol advisor interval ftp 21 6
sscontrol advisor listThis command produces output similar to:
--------------------------------------- | ADVISOR | SITENAME:PORT | TIMEOUT | --------------------------------------- | http | 80 | unlimited | | ftp | 21 | unlimited | ---------------------------------------
sscontrol advisor loglevel http mysite:80 0
sscontrol advisor logsize ftp mysite:21 5000
sscontrol advisor receivetimeout http 80 60
sscontrol advisor report ftp 21This command produces output similar to:
Advisor Report: --------------- Advisor name ............. http Port number .............. 80 sitename ................. mySite Server address ........... 9.67.129.230 Load ..................... 8
sscontrol advisor start ftp 21 ftpadv.log
sscontrol advisor status http 80This command produces output similar to the following:
Advisor Status: --------------- Interval (seconds) ............ 7 Timeout (seconds) ............. Unlimited Connect timeout (seconds).......21 Receive timeout (seconds).......21 Advisor log filename .......... Http_80.log Log level ..................... 1 Maximum log size (bytes) ...... Unlimited
sscontrol advisor stop http 80
sscontrol advisor timeout ftp 21 5
sscontrol advisor version ssl 443
>>-sscontrol--file--+-delete--filename.ext-----------+--------->< +-appendload--filename.ext-------+ +-report-------------------------+ +-save--filename.ext--+--------+-+ | '-force--' | '-newload--filename.ext----------'
The file extension (.ext) can be anything you like and is optional.
Common install directory path -- c:\Program Files\ibm\edge\nd\servers\configurations\component
Native install directory path -- c:\Program Files\ibm\nd\servers\configurations\component
Examples
sscontrol file delete file3 File (file3) was deleted.
sscontrol file newload file1.sv File (file1.sv) was loaded into the Dispatcher.
sscontrol file appendload file2.sv File (file2.sv) was appended to the current configuration and loaded.
sscontrol file report FILE REPORT: file1.save file2.sv file3
sscontrol file save file3 The configuration was saved into file (file3).
>>-sscontrol--help--+-advisor-----+---------------------------->< +-file--------+ +-help--------+ +-host--------+ +-manager-----+ +-metric------+ +-nameserver--+ +-rule--------+ +-server------+ +-set---------+ +-sitename----+ '-status------'
Examples
sscontrol helpThis command produces output similar to:
HELP COMMAND ARGUMENTS: --------------------------------- Usage: help <help option> Example: help name help - print complete help text advisor - help on advisor command file - help on file command host - help on host command manager - help on manager command metric - help on metric command sitename - help on sitename command nameserver - help on nameserver command rule - help on rule command server - help on server command set - help on set command status - help on status commandParameters within < > are variables.
logsize <number of bytes | unlimited> -Set the maximum number of bytes to be logged in the log file
>>-sscontrol--manager--+-interval--seconds-------------------------------+-> +-loglevel--level---------------------------------+ +-logsize--+-size | unlimited-+-------------------+ | '-bytes------------' | +-reach set--+-interval-seconds-----------------+-+ | '-+-loglevel-level---------------+-' | | '---logsize-size | unlimited---' | +-report--sitename+sn2+...+snN--------------------+ +-restart--message--------------------------------+ +-sensitivity--weight-----------------------------+ +-smoothing--smoothing index----------------------+ +-start--+-----------------------+----------------+ | '-logfile--metric_port--' | +-status------------------------------------------+ +-stop--------------------------------------------+ '-version-----------------------------------------' >--------------------------------------------------------------><
The default file is installed in the logs directory. See Appendix F, Sample configuration files. To change the directory where the log files will be kept, see Changing the log file paths.
Examples
sscontrol manager interval 5
sscontrol manager loglevel 0
sscontrol manager logsize 1000000
sscontrol manager reportThis command produces output similar to:
---------------------------------- | SERVER | STATUS | ---------------------------------- | 9.67.129.221| ACTIVE| | 9.67.129.213| ACTIVE| | 9.67.134.223| ACTIVE| ---------------------------------- -------------------------- | MANAGER REPORT LEGEND | -------------------------- | CPU | CPU Load | | MEM | Memory Load | | SYS | System Metric | | NOW | Current Weight | | NEW | New Weight | | WT | Weight | -------------------------- ------------------------------------------------------------------------ | mySite | WEIGHT | CPU 49% | MEM 50% | PORT 1% | SYS 0% | ------------------------------------------------------------------------ | |NOW NEW | WT LOAD | WT LOAD | WT LOAD | WT LOAD | ------------------------------------------------------------------------ | 9.37.56.180 | 10 10 |-99 -1|-99 -1|-99 -1| 0 0| ------------------------------------------------------------------------ | TOTALS:| 10 10 | -1| -1| -1| 0| ------------------------------------------------------------------------ ----------------------------------------- | ADVISOR | SITENAME:PORT | TIMEOUT | ----------------------------------------- | http | 80 | unlimited | -----------------------------------------
sscontrol manager restart Restarting the manager to update codeThis command produces output similar to:
320-14:04:54 Restarting the manager to update code
sscontrol manager sensitivity 10
sscontrol manager smoothing 2.0
sscontrol manager start ndmgr.log
sscontrol manager statusThis command produces output similar to the following example.
Manager status: ============= Metric port................................... 10004 Manager log filename.......................... manager.log Manager log level............................. 1 Maximum manager log size (bytes).............. unlimited Sensitivity level............................. 5 Smoothing index............................... 1.5 Update interval (seconds)..................... 2 Weights refresh cycle......................... 2 Reach log level............................... 1 Maximum reach log size (bytes)................ unlimited Reach update interval (seconds)............... 7
sscontrol manager stop
sscontrol manager version
>>-sscontrol--metric--+-add--sitename+sn2+...+snN:metric+metric1+...+metricN---------------+-> +-remove--sitename+sn2+...+snN:metric+metric1+...+metricN------------+ +-proportions--sitename+sn2+...+snN:proportion1 prop2 prop3...propN--+ '-status--sitename+sn2+...+snN metric+metric1+...+metricN------------' >--------------------------------------------------------------><
Examples
sscontrol metric add site1:metric1
sscontrol metric proportions site1 0 100
sscontrol metric status site1:metric1This command produces output similar to the following:
Metric Status: ------------ sitename ..................... site1 Metric name ................... metric1 Metric proportion ............. 50 Server ......... 9.37.56.100 Metric data .... -1
>>-sscontrol--nameserver--+-start--+-----------------------+-+->< | '-bindaddress--address--' | +-stop-----------------------------+ '-status---------------------------'
>>-sscontrol--rule--+-add--sitename+sn2+...+snN:rule+r2+...+rN--type--value--| value |--| opts |--+-> +-dropserver--sitename+sn2+...+snN:rule+r2+...+rN--server+s2+...+snN----------+ +-remove--sitename+sn2+...+snN:rule+r2+...+rN---------------------------------+ +-set--sitename+sn2+...+snN:rule+r2+...+rN--| value |--| opts |---------------+ +-status--sitename+sn2+...+snN:rule+r2+...+rN---------------------------------+ '-useserver--sitename+sn2+...+snN:rule+r2+...+rN--server+s2+...+snN-----------' >-------------------------------------------------------------->< opts |---+----------------------------------+------------------------| +-beginrange--low--endrange--high--+ +-priority--value------------------+ '-metricname--value----------------'
Examples
sscontrol rule add sitename:rulename type true priority 100
sscontrol rule add sitename:rulename type ip b 9.0.0.0 e 9.255.255.255
sscontrol rule add sitename:rulename type time beginrange 11 endrange 14 sscontrol rule useserver sitename:rulename server05
>>-sscontrol--server--+-add--sitename+sn2+...+snN:server+s2+...+sN--+-------------------------+-+-> | +-metricaddress--address--+ | | '-weight--value-----------' | +-down--sitename+sn2+...+snN:server+s2+...+sN-----------------------------+ +-remove--sitename+sn2+...+snN:server+s2+...+sN---------------------------+ +-set--sitename+sn2+...+snN:server+s2+...+sN--+-------------------------+-+ | +-metricaddress--address--+ | | '-weight--value-----------' | +-status--sitename+sn2+...+snN:server+s2+...+sN---------------------------+ '-up--sitename+sn2+...+snN:server+s2+...+sN-------------------------------' >--------------------------------------------------------------><
Examples
sscontrol server add site1:27.65.89.42
sscontrol server down site1:27.65.89.42
sscontrol server remove :27.65.89.42
sscontrol server up site1:27.65.89.42
>>-sscontrol--set--+-loglevel--level--------+------------------>< +-logsize--+-unlimited-+-+ | '-size------' | '-logstatus--------------'
>>-sscontrol--sitename--+-add--sitename+sn2+...+snN--+-----------------------------------------+-+-> | +-cachelife--value------------------------+ | | +-networkproximity--yes | no--------------+ | | +-proportions--cpu--memory--port--metric--+ | | +-proximitypercentage--value--------------+ | | +-stickytime--time------------------------+ | | +-ttl--time-------------------------------+ | | +-waitforallresponses--yes | no-----------+ | | '-weightbound--weight---------------------' | +-remove--sitename+sn2+...+snN-------------------------------------------+ +-set--sitename+sn2+...+snN--+-----------------------------------------+-+ | +-cachelife--value------------------------+ | | +-networkproximity--yes | no--------------+ | | +-proportions--cpu--memory--port--metric--+ | | +-proximitypercentage--value--------------+ | | +-stickytime--time------------------------+ | | +-ttl--time-------------------------------+ | | +-waitforallresponses--yes | no-----------+ | | '-weightbound--weight---------------------' | '-status--sitename+sn2+...+snN-------------------------------------------' >--------------------------------------------------------------><
Examples
sscontrol sitename add 130.40.52.153
sscontrol sitename set mySite networkproximity yes
sscontrol sitename set mySite cachelife 1900000
sscontrol sitename set mySite proximitypercentage 45
sscontrol sitename set mySite waitforallresponses no
sscontrol sitename set mySite ttl 7
sscontrol sitename set mySite proportions 50 48 1 1
sscontrol sitename remove 130.40.52.153
sscontrol sitename status mySiteThis command produces output similar to:
SiteName Status: --------------- SiteName ........................... mySite WeightBound ........................ 20 TTL ................................ 5 StickyTime ......................... 0 Number of Servers .................. 1 Proportion given to CpuLoad ........ 49 Proportion given to MemLoad ........ 50 Proportion given to Port ........... 1 Proportion given to System metric .. 0 Advisor running on port ............ 80 Using Proximity .................... N
>>-sscontrol--status-------------------------------------------><
Examples
sscontrol statusThis command produces output similar to:
NameServer has been started. Manager has been started. ----------------------------------------- | ADVISOR | SITENAME:PORT | TIMEOUT | ---------------------------------------- | http | 80 | unlimited | -----------------------------------------
This appendix describes how to use the following lbccontrol commands for Consultant for Cisco CSS Switches:
You can enter a minimized version of the ndcontrol command parameters. You only need to enter the unique letters of the parameters. For example, to get help on the file save command, you can enter lbccontrol he f instead of lbccontrol help file.
The "lbc" prefix means load-balancing consultant.
>>-lbccontrol--advisor--+-connecttimeout--name--+-port-----+---timeoutseconds---+-> | '-cluster:port-' | +-interval--name--+-port-----+---seconds----------------+ | '-cluster:port-' | +-list--------------------------------------------------+ +-loglevel--name--+-port-----+---level------------------+ | '-cluster:port-' | +-logsize--+-port-----+---port--+-size | unlimited-+----+ | '-cluster:port-' '-bytes------------' | +-receivetimeout--name--+-port-----+---timeoutseconds---+ | '-cluster:port-' | +-report--name--+-port-----+----------------------------+ | '-cluster:port-' | +-start--name--+-port-----+---+-----------+-------------+ | '-cluster:port-' '-log file--' | +-status--name--+-port-----+----------------------------+ | '-cluster:port-' | +-stop--name--+-port-----+------------------------------+ | '-cluster:port-' | +-timeout--name--+-port-----+---+-seconds | unlimited-+-+ | '-cluster:port-' '-seconds-------------' | '-version--name--+-port-----+---------------------------' '-cluster:port-' >--------------------------------------------------------------><
Advisor Name | Protocol | Port |
---|---|---|
connect | ICMP | 12345 |
db2 | private | 50000 |
ftp | FTP | 21 |
http | HTTP | 80 |
ibmproxy | HTTP (via Caching Proxy) | 80 |
imap | IMAP | 143 |
nntp | NNTP | 119 |
ping | PING | 0 |
pop3 | POP3 | 110 |
smtp | SMTP | 25 |
ssl | SSL | 443 |
telnet | Telnet | 23 |
WLM | private | 10007 |
The default file is advisorname_port.log, for example, http_80.log. To change the directory where the log files will be kept, see Changing the log file paths.
Set the manager proportions to ensure that the advisor information is used.
Examples
lbccontrol advisor connecttimeout http 80 30
lbccontrol advisor interval ftp 21 6
lbccontrol advisor listThis command produces output similar to:
------------------------------ | ADVISOR | PORT | TIMEOUT | ------------------------------ | http | 80 | unlimited | | ftp | 21 | unlimited | ------------------------------
lbccontrol advisor loglevel http 80 0
lbccontrol advisor logsize ftp 21 5000
lbccontrol advisor receivetimeout http 80 60
lbccontrol advisor report ftp 21This command produces output similar to:
Advisor Report: --------------- Advisor name ............. Ftp Port number .............. 21 Cluster address .......... 9.67.131.18 Server address ........... 9.67.129.230 Load ..................... 8 Cluster address .......... 9.67.131.18 Server address ........... 9.67.131.215 Load ..................... -1
lbccontrol advisor start ftp 21 ftpadv.log
lbccontrol advisor status http 80This command produces output similar to the following:
Advisor Status: --------------- Interval (seconds) ............ 15 Timeout (seconds) ............. Unlimited Connect timeout (seconds).......21 Receive timeout (seconds).......21 Advisor log filename .......... Http_80.log Log level ..................... 1 Maximum log size (bytes) ...... Unlimited
lbccontrol advisor stop http 80
lbccontrol advisor timeout ftp 21 5
lbccontrol advisor version ssl 443
>>-lbccontrol--cluster--+-add--cluster+c2+...--+-proportions--active--new--port--system--+-+-> | '-weightbound--weight---------------------' | +-set--cluster+c2+...--+-proportions--active--new--port--system--+-+ | '-weightbound--weight---------------------' | +-remove--cluster--------------------------------------------------+ '-status--cluster--------------------------------------------------' >--------------------------------------------------------------><
Examples
lbccontrol cluster add 130.40.52.153
lbccontrol cluster remove 130.40.52.153
lbccontrol cluster proportions 60 35 5 0
lbccontrol cluster status 9.67.131.167This command produces output similar to
Cluster Status: --------------- Address ................................. 9.67.131.167 Number of target ports .................. 3 Default port weight bound ............... 10 Proportion given to active connections .. 49 Proportion given to new connections ..... 49 Proportion given specific to the port ... 2 Proportion given to system metrics ...... 0
>>-lbccontrol--executor--+-set--+-address--value--------+-+---->< | +-communityname--value--+ | | +-timeout--value--------+ | | '-retries--value--------' | '-status-------------------------'
Examples
lbccontrol executor status Executor Status: ---------------- address ............................. 9.67.131.151 community name ...................... public timeout value ....................... 3 retires value ....................... 2
lbccontrol executor set address 130.40.52.167
>>-lbccontrol--file--+-delete--filename.ext--------------+----->< +-appendload--filename.ext----------+ +-report----------------------------+ +-save--filename.ext--force---------+ '-newload--filename.ext--+--------+-' '-force--'
The file extension (.ext) can be anything you choose and is optional
Common install directory path -- c:\Program Files\ibm\edge\nd\servers\configurations\component
Native install directory path -- c:\Program Files\ibm\nd\servers\configurations\component
Examples
lbccontrol file delete file3 File (file3) was deleted.
lbccontrol file newload file1.sv File (file1.sv) was loaded into the Dispatcher.
lbccontrol file appendload file2.sv File (file2.sv) was appended to the current configuration and loaded.
lbccontrol file report FILE REPORT: file1.save file2.sv file3
lbccontrol file save file3 The configuration was saved into file (file3).
>>-lbccontrol--help--+-advisor---+----------------------------->< +-cluster---+ +-executor--+ +-file------+ +-help------+ +-host------+ +-log-------+ +-manager---+ +-metric----+ +-port------+ +-server----+ +-set-------+ '-status----'
Examples
lbccontrol helpThis command produces output similar to:
HELP COMMAND ARGUMENTS: --------------------------------- Usage: help <help option> Example: help cluster executor - help on executor command cluster - help on cluster command port - help on port command server - help on server command manager - help on manager command metric - help on metric command advisor - help on advisor command file - help on file command host - help on host command log - help on log command set - help on set command status - help on status command help - print complete help textParameters within < > are variables.
>>-lbccontrol--host:--remote_host------------------------------><
lbccontrol host:remote_host
Issue this command at a command prompt, then type any valid lbccontrol command you want issued to the remote Cisco Consultant machine.
>>-lbccontrol--log--+-start-----------------------+------------>< +-stop------------------------+ +-set--+-retention--hours---+-+ | '-interval--seconds--' | '-status----------------------'
>>-lbccontrol--manager--+-interval--seconds-------------------------+-> +-loglevel--level---------------------------+ +-logsize--+-size | unlimited-+-------------+ | '-bytes------------' | +-quiesce--server--+------+-----------------+ | '-now--' | +-reach set--+-interval--seconds----------+-+ | +-loglevel--level------------+ | | '-logsize--size | unlimited--' | +-refresh--refresh cycle--------------------+ +-report--+-----------------+---------------+ | '-cluster+c2+...--' | +-restart--message--------------------------+ +-sensitivity--weight-----------------------+ +-smoothing--value--------------------------+ +-start--+--------------------------+-------+ | '-logfilename--metricport--' | +-status------------------------------------+ +-stop--------------------------------------+ +-unquiesce--server-------------------------+ '-version-----------------------------------' >--------------------------------------------------------------><
The default file is installed in the logs directory. See Appendix F, Sample configuration files. See Changing the log file paths for information on changing the directory where the log files are kept.
Examples
lbccontrol manager interval 5
lbccontrol manager loglevel 0
lbccontrol manager logsize 1000000
lbccontrol manager quiesce 130.40.52.153
lbccontrol manager refresh 3
lbccontrol manager reportThis command produces output similar to:
lbccontrol>>manager report ---------------------------------------- | HOST TABLE LIST | STATUS | ---------------------------------------- | server6| ACTIVE| | server5| ACTIVE| | server4| ACTIVE| | server3| ACTIVE| | server2| ACTIVE| | server1| ACTIVE| ---------------------------------------- ------------------------------------------------------------------------------------------------- | 9.67.154.35 | WEIGHT | ACTIVE % 49 | NEW % 50 | PORT % 1 | SYSTEM % 0 | ------------------------------------------------------------------------------------------------- | PORT: 80 | NOW | NEW | WT | CONNECT | WT | CONNECT | WT | LOAD | WT | LOAD | ------------------------------------------------------------------------------------------------- | server1 | 4 | 4 | 5 | 0 | 5 | 0 | 3| 301| -9999| -1| | server2 | 5 | 5 | 5 | 0 | 5 | 0 | 6| 160| -9999| -1| ------------------------------------------------------------------------------------------------- | PORT TOTALS:| 9 | 9 | | 0 | | 0 | | 461 | | -2 | ------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------- | 9.67.154.35 | WEIGHT | ACTIVE % 49 | NEW % 50 | PORT % 1 | SYSTEM % 0 | ------------------------------------------------------------------------------------------------- | PORT: 443 | NOW | NEW | WT | CONNECT | WT | CONNECT | WT | LOAD | WT | LOAD | ------------------------------------------------------------------------------------------------- | server3 | 4 | 4 | 5 | 0 | 5 | 0 | | 0| -9999| -1| | server4 | 5 | 5 | 5 | 0 | 5 | 0 | 0| 0| -9999| -1| ------------------------------------------------------------------------------------------------- | PORT TOTALS:| 9 | 9 | | 0 | | 0 | | 0 | | -2 | ------------------------------------------------------------------------------------------------- ------------------------------------------------------------------------------------------------- | 9.67.154.34 | WEIGHT | ACTIVE % 49 | NEW % 50 | PORT % 1 | SYSTEM % 0 | ------------------------------------------------------------------------------------------------- | PORT: 80 | NOW | NEW | WT | CONNECT | WT | CONNECT | WT | LOAD | WT | LOAD | ------------------------------------------------------------------------------------------------- | server5 | 5 | 5 | 5 | 0 | 5 | 0 | 5| 160| -9999| -1| | server6 | 0 | 0 | 5 | 0 | 5 | 0 | -9999| -1| -9999| -1| ------------------------------------------------------------------------------------------------- | PORT TOTALS:| 5 | 5 | | 0 | | 0 | | 159 | | -2 | ------------------------------------------------------------------------------------------------- -------------------------------- | ADVISOR | PORT | TIMEOUT | -------------------------------- | http | 80 | unlimited | --------------------------------
lbccontrol manager restart Restarting the manager to update codeThis command produces output similar to:
320-14:04:54 Restarting the manager to update code
lbccontrol manager sensitivity 10
lbccontrol manager smoothing 2.0
lbccontrol manager start ndmgr.log
lbccontrol manager statusThis command produces output similar to the following example.
Manager status: ============= Metric port .................................. 10004 Manager log filename ......................... manager.log Manager log level ............................ 1 Maximum manager log size (bytes) ............. unlimited Sensitivity level ............................ 0.05 Smoothing index .............................. 1.5 Update interval (seconds) .................... 2 Weights refresh cycle ........................ 1 Reach log level .............................. 1 Maximum reach log size (bytes) ............... unlimited Reach update interval (seconds) .............. 7
lbccontrol manager stop
lbccontrol manager version
>>-lbccontrol--metric--+-add--cluster+c2+...+cN:metric+metric1+...+metricN---------------+-> +-remove--cluster+c2+...+cN:metric+metric1+...+metricN------------+ +-proportions--cluster+c2+...+cN:proportion1 prop2 prop3...propN--+ '-status--cluster+c2+...+cN:metric+metric1+...+metricN------------' >--------------------------------------------------------------><
Note: For Cisco Consultant, the cluster address corresponds to the virtual IP (VIP) address of the content rule of the owner in the Cisco CSS Switch configuration.
Examples
lbccontrol metric add 10.10.10.20:metric1
lbccontrol metric proportions 10.10.10.20 48 52
lbccontrol metric status 10.10.10.20:metric1This command produces output similar to the following:
Metric Status: ------------ Cluster ....................... 10.10.10.20 Metric name ................... metric1 Metric proportion ............. 52 Server ......... 9.37.56.100 Metric data .... -1
>>-lbccontrol--port--+-add--cluster:port--+----------------------+-+-> | '-weightbound--weight--' | +-set--cluster:port--+----------------------+-+ | '-weightbound--weight--' | +-remove--cluster:port------------------------+ '-status--cluster:port------------------------' >--------------------------------------------------------------><
Examples
lbccontrol port add 130.40.52.153:80+23
lbccontrol port set 130.40.52.153:80 weightbound 10
lbccontrol port remove 130.40.52.153:23
lbccontrol port status 9.67.131.153:80This command produces output similar to:
Port Status: ------------ Port number .................... 80 Cluster address ................ 9.67.131.153 Number of servers .............. 2 Weight bound ................... 10
>>-lbccontrol--server--+-add--cluster:port:server--+---------------------+------------+-> | +-weight--value-------+ | | +-fixedweight--value--+ | | '-address--address----' | +-set--cluster:port:server--+-weight--value-------+-+ | '-fixedweight--value--' | +-down--cluster:port:server-------------------------+ +-remove--cluster:port:server-----------------------+ +-report--cluster:port:server-----------------------+ +-up--cluster:port:server---------------------------+ '-status--cluster:port:server-----------------------' >--------------------------------------------------------------><
Examples
lbccontrol server add 130.40.52.153:80:27.65.89.42
lbccontrol server remove ::27.65.89.42
lbccontrol server set 130.40.52.153:80:27.65.89.42 weight 10
>>-lbccontrol--set--+-loglevel--level---+---------------------->< +-logsize--+-size-+-+ | '-size-' | '-logstatus---------'
>>-lbccontrol--status------------------------------------------><
Examples
lbccontrol statusThis command produces output similar to:
Manager has been started. -------------------------------- | ADVISOR | PORT | TIMEOUT | -------------------------------- | http | 80 | unlimited | | ftp | 21 | unlimited | --------------------------------
This appendix contains sample configuration files for the Dispatcher component of Network Dispatcher.
Sample files are located in the .../nd/servers/samples/ directory.
#!/bin/ksh
#
# configuration.sample - Sample configuration file for the
Dispatcher component
#
#
# Ensure the root user is the one executing this script.
#
# iam=`whoami`
# if [ "$iam" != "root" ]if [ "$iam" != "root" ]
# then
# echo "You must login as root to run this script"
# exit 2
# fi
#
# First start the server
#
# ndserver start
# sleep 5
#
# Then start the executor
#
# ndcontrol executor start
#
# The Dispatcher can be removed at any time using the
# "ndcontrol executor stop" and "ndserver stop" commands to
# stop the executor and server respectively prior to removing
# the Dispatcher software.
#
# The next step in configuring the Dispatcher is to set the
# NFA (non-forwarding address) and the cluster address(es).
#
# The NFA is used to remotely access the Dispatcher machine
# for administration or configuration purposes. This
# address is required since the Dispatcher will forward packets
# to the cluster address(es).
#
# The CLUSTER address is the hostname (or IP address) to
# which remote clients will connect.
#
# Anywhere in this file, you may use hostnames and IP
# addresses interchangeably.
#
# NFA=hostname.domain.name
# CLUSTER=www.yourcompany.com
# echo "Loading the non-forwarding address"
# ndcontrol executor set nfa $NFA
#
# The next step in configuring the Dispatcher is to create
# a cluster. The Dispatcher will route requests sent to
# the cluster address to the corresponding server machines
# defined to that cluster. You may configure and server
# multiple cluster address using Dispatcher.
# Use a similar configuration for CLUSTER2, CLUSTER3, etc.
#
# echo "Loading first CLUSTER address "
# ndcontrol cluster add $CLUSTER
#
# Now we must define the ports this cluster will use. Any
# requests received by the Dispatcher on a defined port will
# be forwared to the corresponding port of one of the server
# machines.
#
# echo "Creating ports for CLUSTER: $CLUSTER"
# ndcontrol port add $CLUSTER:20+21+80
#
# The last step is to add each of the server machines to the
# ports in this cluster.
# Again, you can use either the hostname or the IP address
# of the server machines.
#
# SERVER1=server1name.domain.name
# SERVER2=server2name.domain.name
# SERVER3=server3name.domain.name
# echo "Adding server machines"
# ndcontrol server add $CLUSTER:20+21+80:
# $SERVER1+$SERVER2+$SERVER3
#
# We will now start the load balancing components of the
# Dispatcher. The main load balancing component is called
# the manager and the second load balancing components are the
# advisors. If the manager and advisors are not running the
# Dispatcher sends requests in a round-robin format. Once the
# manager is started, weighting decisions based on the number
# of new and active connections is employed and incoming
# requests are sent to the best server. The advisors give the
# manager further insight into a servers ability to service
# requests as well as detecting whether a server is up. If
# an advisor detects that a server is down it will be
# marked down (providing the manager proportions have been
# set to include advisor input) and no further requests will be
# routed to the server.
# The last step in setting up the load balancing components
# is to set the manager proportions. The manager updates the
# weight of each of the servers based on four policies:
# 1. The number of active connections on each server.
# 2. The number of new connections to each server.
# 3. Input from the advisors.
# 4. Input from the system level advisor.
# These proportions must add up to 100. As an example, setting
# the manager proportions to
# ndcontrol manager proportions 48 48 0 0
# will give active and new connections 48% input into the
# weighting decision, the advisors will contribute 4% and
# the system input will not be considered.
#
# NOTE: By default the manager proportions are set to 50 50 0 0
#
# echo "Starting the manager..."
# ndcontrol manager start
# echo "Starting the FTP advisor on port 21 ..."
# ndcontrol advisor start ftp 21
# echo "Starting the HTTP advisor on port 80 ..."
# ndcontrol advisor start http 80
# echo "Starting the Telnet advisor on port 23 ..."
# ndcontrol advisor start telnet 23
# echo "Starting the SMTP advisor on port 25 ..."
# ndcontrol advisor start smtp 25
# echo "Starting the POP3 advisor on port 110 ..."
# ndcontrol advisor start pop3 110
# echo "Starting the NNTP advisor on port 119 ..."
# ndcontrol advisor start nntp 119
# echo "Starting the SSL advisor on port 443 ..."
# ndcontrol advisor start ssl 443
#
# echo "Setting the manager proportions..."
# ndcontrol manager proportions 58 40 2 0
#
# The final step in setting up the Dispatcher machine is to
# alias the Network Interface Card (NIC).
#
# NOTE: Do NOT use this command in a high availability
# environment. The go* scripts will configure the NIC and
# loopback as necessary.
# ndcontrol cluster configure $CLUSTER
# If your cluster address is on a different NIC or subnet
from the NFA use the following format for the cluster configure
command.
# ndcontrol cluster configure $CLUSTER tr0 0xfffff800
# where tr0 is your NIC (tr1 for the second token ring card, en0
# for the first ethernet card) and 0xfffff800 is a valid
# subnet mask for your site.
#
#
# The following commands are set to the default values.
# Use these commands as a guide to change from the defaults.
# ndcontrol manager loglevel 1
# ndcontrol manager logsize 1048576
# ndcontrol manager sensitivity 5.000000
# ndcontrol manager interval 2
# ndcontrol manager refresh 2
#
# ndcontrol advisor interval ftp 21 5
# ndcontrol advisor loglevel ftp 21 1
# ndcontrol advisor logsize ftp 21 1048576
# ndcontrol advisor timeout ftp 21 unlimited
# ndcontrol advisor interval telnet 23 5
# ndcontrol advisor loglevel telnet 23 1
# ndcontrol advisor logsize telnet 23 1048576
# ndcontrol advisor timeout telnet 23 unlimited
# ndcontrol advisor interval smtp 25 5
# ndcontrol advisor loglevel smtp 25 1
# ndcontrol advisor logsize smtp 25 1048576
# ndcontrol advisor timeout smtp 25 unlimited
# ndcontrol advisor interval http 80 5
# ndcontrol advisor loglevel http 80 1
# ndcontrol advisor logsize http 80 1048576
# ndcontrol advisor timeout http 80 unlimited
# ndcontrol advisor interval pop3 110 5
# ndcontrol advisor loglevel pop3 110 1
# ndcontrol advisor logsize pop3 110 1048576
# ndcontrol advisor timeout pop3 110 unlimited
# ndcontrol advisor interval nntp 119 5
# ndcontrol advisor loglevel nntp 119 1
# ndcontrol advisor logsize nntp 119 1048576
# ndcontrol advisor timeout nntp 119 unlimited
# ndcontrol advisor interval ssl 443 5
# ndcontrol advisor loglevel ssl 443 1
# ndcontrol advisor logsize ssl 443 1048576
# ndcontrol advisor timeout ssl 443 unlimited
#
The following is a sample Network Dispatcher configuration file called configuration.cmd.sample for use with Window.
@echo off rem configuration.cmd.sample - Sample configuration file for the rem Dispatcher component. rem rem ndserver must be started via Services rem rem rem Then start the executor rem rem call ndcontrol executor start rem rem The next step in configuring the Dispatcher is to set the rem NFA (non-forwarding address) and to set the cluster rem address(es). rem rem The NFA is used to remotely access the Dispatcher rem machine for administration configuration purposes. This rem address is required since the Dispatcher will forward rem packets to the cluster address(es). rem rem The CLUSTER address is the hostname (or IP address) to which rem remote clients will connect. rem rem Anywhere in this file, you may use hostnames and IP rem addresses interchangeably. rem NFA=[non-forwarding address] rem CLUSTER=[your clustername] rem rem set NFA=hostname.domain.name rem set CLUSTER=www.yourcompany.com rem echo "Loading the non-forwarding address" rem call ndcontrol executor set nfa %NFA% rem rem The following commands are set to the default values. rem Use these commands to change the defaults rem call ndcontrol executor set fintimeout 30 rem call ndcontrol executor set fincount 4000 rem rem The next step in configuring the Dispatcher is to create rem a cluster. The Dispatcher will route requests sent to rem the cluster address to the corresponding server machines rem defined to that cluster. You may configure and server rem multiple cluster addresses using Dispatcher. rem Use a similar configuration for CLUSTER2, CLUSTER3, etc. rem rem echo "Loading first CLUSTER address " rem call ndcontrol cluster add %CLUSTER% rem rem Now we must define the ports this cluster will use. Any rem requests received by the Dispatcher on a defined port rem will be forwarded to the corresponding rem port of one of the server machines. rem rem echo "Creating ports for CLUSTER: %CLUSTER%" rem call ndcontrol port add %CLUSTER%:20+21+80 rem rem The last step is to add each of the server machines to rem the ports in this cluster. Again, you can use either the rem hostname or the IP address of the server machines. rem rem set SERVER1=server1name.domain.name rem set SERVER2=server2name.domain.name rem set SERVER3=server3name.domain.name rem echo "Adding server machines" rem call ndcontrol server add %CLUSTER%:20+21+80: rem %SERVER1%+%SERVER2%+%SERVER3% rem rem We will now start the load balancing components of the rem Dispatcher. The main load balancing component is called rem the manager and the second load balancing components are the rem advisors. If the manager and advisors are not rem running the Dispatcher sends requests in a round-robin rem format. Once the manager is started, weighting decisions rem based on the number of new and active connections is rem employed and incoming requests are sent to the best rem server. The advisors give the manager further insight rem into a servers ability to service requests as well as rem detecting whether a server is up. If an advisor detects rem that a server is down it will be marked down (providing the rem manager proportions have been set to include advisor rem input) and no further requests will be routed to the server. rem The last step in setting up the load balancing rem components is to set the manager proportions. The rem manager updates the weight of each of the servers based rem on four policies: rem 1. The number of active connections on each server rem 2. The number of new connections for each server rem 3. Input from the advisors. rem 4. Input from the system level advisor. rem rem These proportions must add up to 100. As an example, rem setting the cluster proportions via rem ndcontrol cluster set <cluster> proportions 48 48 4 0 rem will give active and new connections 48% input into the rem weighting decision, the advisor will contribute 4% and rem the system input will not be considered. rem rem NOTE: By default the manager proportions are set to rem 50 50 0 0 rem echo "Starting the manager..." rem call ndcontrol manager start rem echo "Starting the FTP advisor on port 21 ..." rem call ndcontrol advisor start ftp 21 rem echo "Starting the HTTP advisor on port 80 ..." rem call ndcontrol advisor start http 80 rem echo "Starting the Telnet advisor on port 23 ..." rem call ndcontrol advisor start telnet 23 rem echo "Starting the SMTP advisor on port 25 ..." rem call ndcontrol advisor start smtp 25 rem echo "Starting the POP3 advisor on port 110 ..." rem call ndcontrol advisor start pop3 110 rem echo "Starting the NNTP advisor on port 119 ..." rem call ndcontrol advisor start nntp 119 rem echo "Starting the SSL advisor on port 443 ..." rem call ndcontrol advisor start ssl 443 rem rem echo "Setting the cluster proportions..." rem call ndcontrol cluster set %CLUSTER% proportions 58 40 2 0 rem rem The final step in setting up the Dispatcher machine is rem to alias the Network Interface Card (NIC). rem rem NOTE: Do NOT use this command in a high availability rem environment. The go* scripts will configure the NIC and rem loopback as necessary. rem rem ndcontrol cluster configure %CLUSTER% rem If your cluster address is on a different NIC or subnet rem from the NFA use the following format for the cluster rem configure command. rem ndcontrol cluster configure %CLUSTER% tr0 0xfffff800 rem where tr0 is your NIC (tr1 for the second token ring card, rem en0 for the first ethernet card) and 0xfffff800 is rem a valid subnet mask for your site. rem rem rem The following commands are set to the default values. rem Use these commands to guide to change from the defaults. rem call ndcontrol manager loglevel 1 rem call ndcontrol manager logsize 1048576 rem call ndcontrol manager sensitivity 5.000000 rem call ndcontrol manager interval 2 rem call ndcontrol manager refresh 2 rem rem call ndcontrol advisor interval ftp 21 5 rem call ndcontrol advisor loglevel ftp 21 1 rem call ndcontrol advisor logsize ftp 21 1048576 rem call ndcontrol advisor timeout ftp 21 unlimited rem call ndcontrol advisor interval telnet 23 5 rem call ndcontrol advisor loglevel telnet 23 1 rem call ndcontrol advisor logsize telnet 23 1048576 rem call ndcontrol advisor timeout telnet 23 unlimited rem call ndcontrol advisor interval smtp 25 5 rem call ndcontrol advisor loglevel smtp 25 1 rem call ndcontrol advisor logsize smtp 25 1048576 rem call ndcontrol advisor timeout smtp 25 unlimited rem call ndcontrol advisor interval http 80 5 rem call ndcontrol advisor loglevel http 80 1 rem call ndcontrol advisor logsize http 80 1048576 rem call ndcontrol advisor timeout http 80 unlimited rem call ndcontrol advisor interval pop3 110 5 rem call ndcontrol advisor loglevel pop3 110 1 rem call ndcontrol advisor logsize pop3 110 1048576 rem call ndcontrol advisor timeout pop3 110 unlimited rem call ndcontrol advisor interval nntp 119 5 rem call ndcontrol advisor loglevel nntp 119 1 rem call ndcontrol advisor logsize nntp 119 1048576 rem call ndcontrol advisor timeout nntp 119 unlimited rem call ndcontrol advisor interval ssl 443 5 rem call ndcontrol advisor loglevel ssl 443 1 rem call ndcontrol advisor logsize ssl 443 1048576 rem call ndcontrol advisor timeout ssl 443 unlimited rem
The following is a sample advisor file called ADV_sample.
/**
* ADV_sample: The Network Dispatcher HTTP advisor
*
*
* This class defines a sample custom advisor for Network Dispatcher.
* Like all advisors, this custom advisor extends the function of the
* advisor base, called ADV_Base. It is the advisor base that actually
* performs most of the advisor's functions, such as reporting loads back
* to the Network Dispatcher for use in the Network Dispatcher's weight
* algorithm. The advisor base also performs socket connect and close
* operations and provides send and receive methods for use by the advisor.
* The advisor itself is used only for sending and receiving data to and
* from the port on the server being advised.
* The TCP methods within the advisor base are timed to calculate the load.
* A flag within the constructor in the ADV_base
* overwrites the existing load with the new load returned from the advisor
* if desired.
*
* Note: Based on a value set in the constructor, the advisor base supplies
* the load to the weight algorithm at specified intervals. If the actual
* advisor has not completed so that it can return a valid load, the advisor
* base uses the previous load.
*
* NAMING
*
* The naming convention is as follows:
*
* - The file must be located in the following Network Dispatcher
* Directories:
*
* nd/servers/lib/CustomAdvisors/
* (nd\servers\lib\CustomAdvisors on Windows 2000)
*
* - The Advisor name must be preceded with "ADV_". The advisor can
* be started with only the name, however; for instance, the "ADV_sample"
* advisor can be started with "sample".
*
* - The advisor name must be in lowercase.
*
* With these rules in mind, therefore, this sample is referred to as:
*
* <base directory>/lib/CustomAdvisors/ADV_sample.class
*
*
* Advisors, as with the rest of Network Dispatcher, must be compiled with
* the prereq version of Java.
* To ensure access to Network Dispatcher classes, make sure that the
* ibmnd.jar file (located in the lib subdirectory of the base directory)
* is included in the system's CLASSPATH.
*
*
* Methods provided by ADV_Base:
*
* - ADV_Base (Constructor):
*
* - Parms
* - String sName = Name of the advisor
* - String sVersion = Version of the advisor
* - int iDefaultPort = Default port number to advise on
* - int iInterval = Interval on which to advise on the servers
* - String sDefaultLogFileName = Unused. Must be passed in as "".
* - boolean replace = True - replace the load value being calculated
* by the advisor base
* False - add to the load value being calculated
* by the advisor base
* - Return
* - Constructors do not have return values.
*
* Because the advisor base is thread based, it has several other methods
* available for use by an advisor. These methods can be referenced using
* the CALLER parameter passed in getLoad().
*
* These methods are as follows:
*
* - send - Send a packet of information on the established socket
* connection to the server on the specified port.
* - Parms
* - String sDataString - The data to be sent is sent in the form of a
* string
* - Return
* - int RC - Whether the data was sucessfully sent or not: zero
* indicates data was sent; a negative integer indicates an
* error.
*
* - receive - Receive information from the socket connection.
* - Parms
* - StringBuffer sbDataBuffer - The data received during the receive
* call
* - Return
* - int RC - Whether the data was successfully received or not; zero
* indicates data was sent; a negative integer indicates an error.
*
* If the function provided by the advisor base is
* not sufficient, you can create the appropriate function within the
* advisor and the methods provided by the advisor base will then be
* ignored.
*
* An important question regarding
* the load returned is whether to apply it to the load being generated
* within the advisor base, or to replace it; there are valid instances of
* both situations.
*
* This sample is essentially the Network Dispatcher HTTP advisor. It
* functions very simply:
* a send request--an http head request--is issued. Once a response is
* received, the getLoad method terminates, flagging the advisor base to
* stop timing the request. The method is then complete. The information
* returned is not parsed; the load is based on the time required
* to perform the send and receive operations.
*/
package CustomAdvisors;
import com.ibm.internet.nd.advisors.*;
public class ADV_sample extends ADV_Base implements ADV_MethodInterface
{
String COPYRIGHT = "(C) Copyright IBM Corporation 1997,
All Rights Reserved.\n";
static final String ADV_NAME = "Sample";
static final int ADV_DEF_ADV_ON_PORT = 80;
static final int ADV_DEF_INTERVAL = 7;
// Note: Most server protocols require a carriage return ("\r") and line
// feed ("\n") at the end of messages. If so, include them in your
// string here.
static final String ADV_SEND_REQUEST =
"HEAD / HTTP/1.0\r\nAccept: */*\r\nUser-Agent: " +
"IBM_Network_Dispatcher_HTTP_Advisor\r\n\r\n";
/**
* Constructor.
*
* Parms: None; but the constructor for ADV_Base has several parameters
* that must be passed to it.
*
*/
public ADV_sample()
{
super( ADV_NAME,
"2.0.0.0-03.27.98",
ADV_DEF_ADV_ON_PORT,
ADV_DEF_INTERVAL,
"", // not used
false);
super.setAdvisor( this );
}
/**
* ADV_AdvisorInitialize
*
* Any Advisor-specific initialization that must take place after the
* advisor base is started.
* This method is called only once and is typically not used.
*/
public void ADV_AdvisorInitialize()
{
return;
}
/**
* getLoad()
*
* This method is called by the advisor base to complete the advisor's
* operation, based on details specific to the protocol. In this sample
* advisor, only a single send and receive are necessary; if more complex
* logic is necessary, multiple sends and receives can be issued.
* For example, a response might be received and parsed. Based on the
* information learned thereby, another send and receive could be issued.
*
* Parameters:
*
* - iConnectTime - The current load as it refers to the length of time it
* took to complete the connection to the server through
* the specified port.
*
* - caller - A reference to the advisor base class where the Network
* Dispatcher-supplied methods are to perform simple TCP
* requests, mainly send and receive.
*
* Results:
*
* - The load - A value, expressed in milliseconds, that can either be
* added to the existing load, or that can replace the existing load,
* as determined by the constructor's "replace" flag.
*
* The larger the load, the longer it took the server to respond;
* therefore, the higher the weight will be within Network Dispatcher
* regarding load balancing.
*
* If the value is negative, an error is assumed. An error from an
* advisor indicates that the server the advisor is trying to reach is
* not accessible and has been identified as being down.
* Network Dispatcher will not attempt to load balance to a server that
* is down. Network Dispatcher will resume load balancing to the server
* when a positive value is received.
*
* A value of zero is typically not returned; Network Dispatcher handles
* a load of zero in a special way. Zero is assumed to indicate an
* unknown status, and Network Dispatcher gives the server a high
* weight in response.
*/
public int getLoad(int iConnectTime, ADV_Thread caller)
{
int iRc;
int iLoad = ADV_HOST_INACCESSIBLE; // -1
// Send tcp request
iRc = caller.send(ADV_SEND_REQUEST);
if (iRc >= 0)
{
// Perform a receive
StringBuffer sbReceiveData = new StringBuffer("");
iRc = caller.receive(sbReceiveData);
// If the receive is successful, a load of zero is returned.
// This is because the "replace" flag is set to false,
// indicating that the load built within the base advisor is
// to be used.
// Since nothing was done with the returned data, additional
// load is not necessary.
// Note: it is known that the advisor base load will not be
// zero, therefore a zero load will
// not be returned for use in calculating the weight.
if (iRc >= 0)
{
iLoad = 0;
}
}
return iLoad;
}
} // End - ADV_sample
This appendix describes how to set up a 2-tier, high availability configuration combining the capabilities of two Network Dispatcher components (the Dispatcher component and the CBR component) along with Caching Proxy.
The server machine set up for Figure 30 is the following:
Figure 30 shows a basic representation of multiple servers (EdgeServer1, EdgeServer2, EdgeServer3) load balancing across multiple backend Web servers. The CBR component uses Caching Proxy to forward requests based on the content of the URL to the backend Web servers. The Dispatcher component is used to load balance the CBR components across the Edge Servers. The high availability feature of the Dispatcher component is used to ensure that requests to the backend servers continue even if the primary high availability machine (EdgeServer1) fails at any time.
Basic Configuration Guidelines:
Sample lines referred to in notes 1-4 (above):
Hostname www.company.com SendRevProxyName yes Caching ON CacheMemory 128000 K Proxy /* http://www.company.com/* www.company.com
Sample Configuration Files:
The following sample configuration files are similar to files that are created when setting up an Edge Server configuration as shown in Figure 30. The sample configuration files represent the files for the Dispatcher and CBR components of Network Dispatcher. In the sample configuration, a single ethernet adapter is used for each of the Edge Server machines and all addresses are represented within a private subnet. The sample configuration files use the following IP addresses for the specified machines:
Sample Configuration file for Dispatcher component on Primary high availability Edge Server:
ndcontrol executor start ndcontrol cluster add 192.168.1.11 primaryhost 192.168.1.10 ndcontrol port add 192.168.1.11:80 ndcontrol server add 192.168.1.11:80:edgeserver1 address 192.168.1.10 ndcontrol server add 192.168.1.11:80:edgeserver2 address 192.168.1.20 ndcontrol server add 192.168.1.11:80:edgeserver3 address 192.168.1.30 ndcontrol manager start manager.log 10004 ndcontrol highavailability heartbeat add 192.168.1.10 192.168.1.20 ndcontrol highavailability backup add primary auto 4567
Sample Configuration file for CBR component on the Edge Servers:
cbrcontrol set loglevel 1 cbrcontrol executor start cbrcontrol cluster add 192.168.1.11 cbrcontrol port add 192.168.1.11:80 cbrcontrol server add 192.168.1.11:80:webserverA address 192.168.1.71 cbrcontrol server add 192.168.1.11:80:webserverB address 192.168.1.72 cbrcontrol server add 192.168.1.11:80:webserverC address 192.168.1.73 cbrcontrol rule add 192.168.1.11:80:webA_rule type content pattern (URI=*WSA*)|(URI=*wsA*) priority 21 cbrcontrol rule useserver 192.168.1.11:80:webA_rule webserverA cbrcontrol rule add 192.168.1.11:80:webB_rule type content pattern (URI=/WS_B*) priority 22 cbrcontrol rule useserver 192.168.1.11:80:webB_rule webserverB cbrcontrol rule add 192.168.1.11:80:webC_rule type content pattern URI=*webC* priority 23 cbrcontrol rule useserver 192.168.1.21:80:webC_rule webserverC
In many circumstances, you can use keys or key combinations to perform operations that can also be done through mouse actions. Many menu actions can be initiated from the keyboard.
Consult the documentation for your operating system for instructions on using the keyboard.
Network Dispatcher includes an online help facility, which describes the tasks you will perform while installing, planning, configuring, and operating the product.
To get help for the current window, click the question mark ( ? ) in the upper right corner. Choose from:
For additional information on using Network Dispatcher, refer to:
http://www.ibm.com/software/webservers/edgeserver
http://www.ibm.com/software/webservers/edgeserver/support.html
Click on Search for Network Dispatcher hints and tips.
References in this publication to IBM products, programs, or services do not imply that IBM intends to make them available in all countries in which IBM operates. Any reference to an IBM product, program, or service is not intended to state or imply that IBM product, program or service may be used. Subject to IBM's valid intellectual property or other legally protectable rights, any functionally equivalent product, program, or service may be used instead of the IBM product, program, or service. The evaluation and verification of operation in conjunction with other products, except those expressly designated by IBM, are the responsibility of the user.
IBM may have patents or pending patent applications covering subject matter described in this document. The furnishing of this document does not give you any license to these patents. You can send license inquiries, in writing, to the IBM Director of Licensing, IBM Corporation, North Castle Drive, Armonk, NY 10504-1785, U.S.A.
Licensees of this program who wish to have information about it for the
purpose of enabling: (i) the exchange of information between
independently created programs and other programs (including this one) and
(ii) the mutual use of the information which has been exchanged, should
contact:
Site Counsel
IBM Corporation
P.O. Box 12195
3039 Cornwallis Avenue
Research Triangle Park, NC 27709-2195
USA
The licensed program described in this document and all licensed material available for it are provided by IBM under terms of the IBM Customer Agreement.
This document is not intended for production use and is furnished as is without any warranty of any kind, and all warranties are hereby disclaimed including the warranties of merchantability and fitness for a particular purpose.
This product includes computer software created and made available by CERN. This acknowledgement shall be mentioned in full in any product which includes the CERN computer software included herein or parts thereof.
The following terms are registered trademarks or trademarks of IBM Corporation in the United States, other countries or both.
AIX
IBM
IBMLink
LoadLeveler
OS/2
NetView
WebSphere
Lotus is a registered trademark of Lotus Development Corporation in the United States, other countries or both.
Domino is a trademark of Lotus Development Corporation in the United States, other countries or both.
Tivoli is a registered trademark of Tivoli Systems, Inc in the United States, other countries, or both.
Java and all Java-based trademarks and logos are trademarks or registered trademarks of Sun Microsystems, Inc. in the United States, other countries or both.
Solaris is a trademark of Sun Microsystems, Inc.in the United States, other countries, or both.
Microsoft and Windows 2000 are trademarks or registered trademarks of the Microsoft Corporation in the United States, other countries, or both.
Cisco is a registered trademark of Cisco Systems, Inc. in the United States and certain other countries.
HP is a trademark of the Hewlett-Packard Company in the United States, other countries, or both.
Linux is a registered trademark of Linus Torvalds.
Red Hat is a registered trademark of Red Hat, Inc.
UNIX is a registered trademark of The Open Group in the United States and other countries.
Other company, product, and service names, which may be denoted by a double asterisk (**), may be trademarks or service marks of others.