Chapter 13
After the Installation

This chapter describes steps that should be completed after installing software through ptx/INSTALL. Refer to the product release notes and administration guides for information about setting up and configuring the software. Most administration guides are available online; refer to the HTML documentation CD shipped with DYNIX/ptx V4.4.10.


Remove Unneeded Installation Directories

Remove the newfiles.<date-time> and oldfiles.<date-time> directories from each /usr/options subdirectory. These directories use large amounts of disk space and are no longer needed after the installation is complete.


Sync Files Between the Original Root Disk and the Alternate Disk

This optional procedure applies only to INIT ALT DISK DELTA and ALT DISK DELTA installations.

The Alternate Disk Compare menu option locates any file modifications that were made on the original root disk after the image of the operating system was created on an alternate disk. This comparison should be done after booting on the alternate disk.

The Alternate Disk Compare menu option does a three-way comparison among the following files:

The comparison creates three files: NewFiles, ClobberFiles, and ConflictFiles. The files are located in the directory /usr/options/base/altdelta.

To perform the Alternate Disk Compare, complete the following steps:

  1. Mount the original root filesystem under another name such as /mnt:

    # /etc/devbuild vtoc /dev/rdsk/sd0
    # /etc/mount /dev/dsk/s0s0 /mnt
  2. Start the menu system and select the Alternate Disk Compare option from the Software Management menu.

       
      Get Files That Changed During Alternate Disk Delta
     
    Enter mountpoint for the root filesystem     ___________
    on the original root disk (e.g., /mnt).
    If the corresponding /usr directory is
    on a separate partition, that partition
    must be mounted. The alternate-disk
    mountpoint must be in the root directory
    (/mnt is valid;/a/mnt is not).
     
    Enter <y> if you want to overwrite          ____________
    current root disk files with New and
    Changed files from the original root 
    disk after reviewing the lists. No files
    are saved. Conflict files must be
    resolved by hand. Enter <n> to only 
    generate a list of New, Changed, and
    Conflict files; no files are overwritten.
    

  3. The first prompt asks for the mount point for the root filesystem on the original root disk. Enter the mountpoint that you specified in step 1. (The original root filesystem must be mounted.)

  4. The second prompt allows you to specify whether you want files on the new root filesystem (the filesystem on the alternate disk) to be automatically overwritten by the revised files from the original root filesystem.

  5. When you select Run to execute the form, the system will generate the files NewFiles, ClobberFiles, and ConflictFiles.

    If you answered N to the second prompt on the form, the files specified in NewFiles, ClobberFiles, and ConflictFiles will not be copied to the new root disk. If you want to copy the files manually, use the following commands.

    $ cd /mnt
    $ cpio -pdmau / < /usr/options/base/altdelta/NewFiles
    $ cpio -pdmau / < /usr/options/base/altdelta/ClobberFiles
    $ cpio -pdmau / < /usr/options/base/altdelta/ConflictFiles

    If you answered Y to the second prompt on the form, you can view or edit the contents ofNewFiles and ClobberFiles before any files are copied. You will be prompted to press any key; when you do this, the vi editor is invoked on the files. When you exit vi, the following prompt appears:

    Do you still wish to overwrite / files with the /mnt file list (y/n)?
    

    If you answer Y, the Alternate Disk Compare procedure will copy the files specified inNewFiles and ClobberFiles to the new root disk, overwriting the files currently on the disk. Files in conflict must be viewed or copied manually. Use the cpio command shown above to copy the files.

    If you answer N, you will need to view or copy the files manually. Use the cpio commands shown earlier to copy the files.


Create a Custom Miniroot

After all software is installed on your system, it is strongly recommended that you create a custom miniroot. You will need to use this miniroot if it should become necessary to restore the root filesystem. See the DYNIX/ptx V4.4 System Recovery and Troubleshooting Guide for information about creating a miniroot.


Synchronize System Clocks

To ensure that problems can be diagnosed quickly, we recommend that you synchronize all time clocks on your system. If you do not do this, diagnosing certain problems can take longer because time skew between the clocks can make it difficult to correlate event entries in the system logs.

To synchronize the clocks on hosts running DYNIX/ptx, configure either timed or xntpd as described in the ptx/TCP/IP Administration Guide.


Product-Specific Tasks


ptx/CLUSTERS

If you installed ptx/CLUSTERS and you were not prompted for cluster parameters, you must manually configure required clustered parameters. To do so, use either ptx/ADMIN (Cluster Administration->Cluster Operations->Change Cluster Parameters) or the clustadm -P command to set the following cluster parameters:

Refer to Chapter 2, "Administering and Monitoring an Existing Cluster," of the ptx/CLUSTERS Administration Guide or the clustadm(1M) man page for details.


ptx/CTC

Because of changes in ptx/SVM and ptx/CLUSTERS, you may need to modify your ptx/CTC configuration. Read the ptx/CTC Release Notes for instructions on updating the path to the ptx/CTC database and modifying ptx/CTC failover scripts.

ptx/CLUSTERS and ptx/CTC should always be installed together. If an updated version of ptx/CLUSTERS is installed and ptx/CTC is not reinstalled, the ptx/CTC menus in ptx/ADMIN will be removed. To correct this problem, reinstall ptx/CTC from the distribution CD-ROM and the menus will reappear.


ptx/NFS

Once installed, ptx/NFS cannot be deinstalled.


ptx/RAID

The Licensed Internal Code (LIC) microcode needed for the Disk Array Storage System (DASS) is installed with ptx/RAID. Refer to the ptx/RAID Release Notes for information about downloading the LIC.

Also verify that the scsidisk_standard_timeout parameter is set correctly. The parameter is in the /usr/conf/uts/io/scsidisk/scsidisk_space.c file and should be set as follows:

 int scsidisk_standard_timeout = 150;  /* seconds (was 35) */

ptx/SNA

To complete the installation, configure the system as described in the ptx/SNA PU 2.1 Base Server Release Notes. You must also obtain from Customer Support an Encoded Option String [EOS] that is valid for your system.


ptx/SPDRIVERS


Associate Tape Drives With Media Changers By Using the mcbuild Utility

When using directly-connected devices, you should use the mcbuild utility to associate tape drives with each media changer or library device. You can create or update these associations at any time-a system reboot is not required. This task does not apply to ACSLS-connected libraries.

When mcbuild is not used, a media changer driver tries to determine the names of its connected tape drives the first time the media changer device is opened. The media changer driver then keeps those tape drive names until the system is rebooted. This method works most of the time, especially when there is only one media changer on the host and all the tape drives connected to the host are in the media changer. However, because the media changer device cannot return unique identifiers for the tape drives connected to it, the media changer driver has no way to positively associate a tape drive with that media changer device. Also, if you rename a tape device with the devctl command, the output from the mc command does not reflect the current device configuration until the system is rebooted.

To prevent these problems, you should use the mcbuild utility to explicitly associate tape drives with every media changer on the host. Even if you only have one media changer on the host, it is a good system administration practice to use the mcbuild utility.

To use the mcbuild utility to associate tape drives to media changer devices, complete the following steps:

  1. Create or edit the /etc/mctab file to contain one line for each media changer device and its associated tape drives.


    ATTENTION

    Any time you add, remove, or rename media changers or the tape drives in these media changers, you must update the /etc/mctab file to reflect the current device configuration. Otherwise, the output from the mc command will not reflect the changes.


    Each line in the /etc/mctab file should have this format:

    mc_device_name drive1 drive2 drive3 ...

    The order given for the tape drives must correspond to their order as storage elements in the library. This is important because the naming order of tape drives may not correspond directly to their order in a library. For example, a library may have four tape drives with SCSI ID's 0, 1, 2, and 3. Because of the way they are cabled to the system, the system may probe them in a different order (such as 2, 3, 0, 1). The system will assign drive names in probe order, but mc(1) will erroneously assume that the drive name sequence follows its own ordering. mcbuild must be used to rectify this situation by listing the drive names in the correct order with respect to the library.

    For example, to associate the tape drives tl0 through tl3 with the HP DLT4000 library /dev/mch/ml0 and to associate the tape drives tl4 through tl8 with the STK 9710 library /dev/mch/mw0, add the following lines to the /etc/mctab file:

    /dev/mch/ml0 tl0 tl1 tl2 tl3 tl4
    /dev/mch/mw0 tl5 tl6 tl7 tl8 tl9

    If you do not know which tape drives belong to which library, you can determine this information as follows:

  2. Run the mcbuild utility to associate the tape drives and media changers that are defined in the /etc/mctab file.

    # mcbuild

Verify Hardware Detection

After installing ptx/SPDRIVERS software and appropriate hardware, you can verify that the software is properly detecting the hardware. Complete the following steps:

  1. Use the dumpconf command to verify that the appropriate devices were found at system boot. Following is sample output for td devices:

    # /etc/dumpconf -d | grep td
    td0 td 0 0x00000000 S scsi scsibus2
    td1 td 1 0x00000010 S scsi scsibus2
    td2 td 2 0x00000010 S scsi scsibus17
    td3 td 3 0x00000020 S scsi scsibus17
    td4 td 4 0x00000030 S scsi scsibus17
    td5 td 5 0x00000040 S scsi scsibus17
  2. Use the dumpconf command to verify that the appropriate media changers (libraries) were found at system boot. The following output is for ml and mx devices:

    # /etc/dumpconf -d | egrep "ml | mx"
    mx0 mx 0 0x00000040 S scsi scsibus3
    ml0 ml 0 0x00000050 S scsi scsibus9
  3. Run the infodev -a command on each device to verify that the output returned from the SCSI inquiry command is as expected. (There must be a tape in the drive when you run this command.) For example, to display information for /dev/rmt/tl0, enter the following command:

    # infodev -a /dev/rmt/tl0
  4. At this point, you may want to verify the operation of each library or tape drive. To do this, load a tape into each drive, write data to the tape, read data from the tape back to the system, and verify the write/read operation. You should use tar or cpio to perform these basic operational tests.


ptx/SVM

To complete the installation, follow the procedure entitled "Start the ptx/SVM Configuration Daemon" in Chapter 2 of the ptx/SVM Administration Guide. Then read the ptx/SVM Administration Guide for information about laying out your ptx/SVM system, placing disks under ptx/SVM control, and creating disk groups and objects.


ptx/SYNC

To complete the installation, refer to the ptx/SYNC Release Notes and continue with the procedure corresponding to your system's architecture.


ptx/TCP/IP

Ensure that you have your original version of the /etc/inetd.conf file. You may also need to make the following change to this file.

The version of reshd distributed with this release requires a service type of stream. Locate the entry for shell in the /etc/inetd.conf file. If the entry specifies the tli service type, as in the following example, you will need to change it to stream.

shell   tli  tcp   nowait  root  /usr/etc/reshd  reshd -R

The corrected entry is as follows:

shell   stream  tcp   nowait  root  /usr/etc/reshd  reshd -R

If you do not make this change, remote resh services will not function.


ptx/X.25

If you saved your previous ptx/X.25 configuration, complete the following steps to restore that configuration.


ATTENTION

The configuration files supplied with the newly installed version of ptx/X.25 will be overwritten by the saved configuration files. If you want to save the new configuration files before they are overwritten, follow the procedure in the section "Updates of ptx/X.25" in Chapter 2.


  1. As root, enter the following command, where directory is the directory in which you saved the configuration.

    # /usr/lib/x25/rstx25conf directory
  2. When the following message appears, answer y to confirm that you want to restore the saved X.25 configuration.

           WARNING         WARNING         WARNING
    Restoring X.25 configuration will overwrite existing files.
    It will overwrite some files even if you are not using them
    currently.
    Are you sure you want to restore X.25 configuration? (y/n)
    

  3. If your /usr/lib/x25/x25d.action file contains a reference to xlogind, the following message appears:

    Please edit the file /usr/lib/x25/x25d.action to change 
    /usr/etc/xlogind to /bin/login

    You may want to edit this file, because the xlogind program is no longer used in the ptx/X.25 product.

  4. When the X.25 configuration files have been restored, the following message appears:

    Saved X.25 configuration has been restored
  5. If the HDLC Baud Rate parameter is set to a non-zero value in any of the restored configuration files, the following message appears:

    The Level II parameter HDLC Baud Rate MUST be set to 0 if
    clock is provided by an external source.  This parameter
    must be set correctly for the X.25 network to operate
    The following parameter files have this parameter set to
    a non-zero value.  If you use them, be sure to check this
    parameter's value.
    

    List of files

    A non-zero HDLC baud rate indicates that clocking will be provided by an internal source. Internal clocking is supported only on the SCC-I board. If you are using ptx/X.25 with an SCC-V or DCP 8804 board, the HDLC baud rate must be set to 0. You can change this parameter on the Other Level 2 Parameters form during ptx/X.25 configuration.

  6. When the following prompt appears, press Return, then y and Return:

    The format of parameter files has changed for this release
    of ptx/X.25.
    The following parameter files are not valid
    

    List of files
    .
    .
    EOF (Press <Return> here)
    Would you like me to fix ALL of them? (y/n)

    This will convert the restored configuration files to the new format. Note that if you do not convert these files, you cannot use them with this release of ptx/X.25.

    Ignore error messages with the following form:

    /usr/lib/x25/x25fixfile: error in file filename for parameter use_d_bit
  7. After the saved configuration files have been restored, you can remove the saved directory structure. Before removing the directory, ensure that there are no files in the directory that you wish to save. Then enter the following command, wheredirectory is the directory in which you saved the configuration.

    # rm -r directory

ptx/XWM


Restart xdm

If you rebooted the operating system after installing ptx/XWM, the system will restart xdm automatically. If you did not reboot, the previous version of xdm will still be running. To start the new version of xdm, complete the following steps:

  1. Kill the running xdm process:

    # kill -9 `cat /usr/lib/X11/xdm/xdm-pid`

  2. Restart xdm.


Customize .Xresources

If you saved a customized version of the default .Xdefaults file, merge your modifications into the new /usr/admin.ptx/etc/base.dir/.Xresources file.


Use the Font Server


ATTENTION

The name of the font server binary was changed from fs to xfs in ptx/XWM V4.5.0.


If you plan to use the font server, edit the following line in the /usr/lib/X11/fs/config file so that the ordered list of font path elements can be accessed from the host running xfs.

catalogue =

This font path is often the same as the font path used by X terminals at your site. For more information, refer to the xfs(1) man page and to the font server information in O'Reilly, X Window System Administrator's Guide, Volume 8.


Link X11R5 Libraries Dynamically

The X11R5 libraries must be linked dynamically. If you have an application that must link with the X11R5 libraries in the directory /usr/X11R5/lib, and must also link statically (using the compiler option -Bstatic) with other libraries, you will need to specify -Bdynamic to link the X11R5 libraries.

To link the X11R5 libraries, you must either specify a cc command line option (-L/usr/X11R5/lib) or set an environment variable as described in the ptx/XWM Release Notes. In addition, either allow the dynamic libraries to be linked by default, or, if any other libraries are linked statically with -Bstatic, use -Bdynamic before linking the Motif and X11 libraries. Following are some examples.


QCIC

System configurations can exist in which too much disk bandwidth is allocated to individual SCSI buses. A Pbay fully populated with 12 high speed drives (e.g., 9 GB or 18 GB IBM), deep device queues (i.e., greater than two), and multiple high priority initiators would be an example of such a non-optimal setup. The ping-poll starvation management algorithm is tuned for a heavily loaded, but not a severely overloaded, SCSI bus. If you need to run a non-optimal configuration like this example, you may need to tune certain bus utilization parameters to avoid starvation failures. See the section "SCSI Starvation (248319)" in the QCIC Software V3.4.1 Release Notes for more information.