Client logs and traces

Use the logging and tracing functions that are provided on each SAN File System client to help diagnose client-related problems. There is also a client dump facility to help with data collection in the event that IBM® Support personnel requests this information.

Windows client logs, traces, and dumps

For Windows users, client log messages are written to the standard event log. To view the log, click Start –> Programs –> Administrative Tools –> Event Viewer.

Windows client tracing is written in the file C:\Program Files\IBM\Storage Tank\Client\log\sanfs.log for Windows. You must first enable tracing, using stlog command. You can view the trace log using a standard text editor.

You can configure Windows to generate a dump file if the Windows SAN File System client ends abnormally. By default, the file name is C:\WINNT\memory.dmp.

AIX client logs, traces, and dumps

In AIX, use the stfsdebug command and the syslog utility to enable tracing and logging on an AIX client. This client generates both log and trace messages, which are routed through the syslog utility on the AIX operating system. The syslog utility captures log and trace output from the kernel, as well as other operating system services. By default, the syslog utility discards all kernel output. However, you can configure the syslog utility to specify a destination for the messages by modifying the/etc/syslog.conf file. You use the stfsdebug command to enable tracing for an AIX client. In addition, you can specify which components (called classes) are traced, as well as the level of detail to include. You can also use stfsdebug to query the current status of all trace classes. Trace output enabled stfsdebug is sent to the syslog facility.

You can initiate a kernel dump of an AIX system if is still running but no longer responding to commands. Initiate a kernel dump using the sysdumpstart command from a system prompt using telnet or Secure Shell (SSH) tools; if you cannot establish remote access to the client, press the system reset button.

Linux client traces

To turn on tracing for the Linux client:
  1. Verify that there is an entry in /etc/syslog.conf file for kern.debug (for example, kern.debug /var/log/kernmsg).
  2. After the client module is loaded, edit the proc entries:
    1. Edit the "/proc/fs/stfs/debug" entry turn on tracing for the needed components (for example, change OFF to ON).
    2. Edit the "/proc/fs/stfs/csmdebuglevel" entry to set the CSM tracing level.

The trace output is placed in the location specified in the /etc/syslog.conf file (for example, /var/log/kernmsg).

Solaris client logs, traces, and dumps

You use the sanfs_ctl utility to enable tracing on a Solaris client. Logging and tracing output is stored in the standard system log in file /var/adm/messages.

The Solaris operating system automatically creates a disk copy of the physical memory when a system crash occurs. This disk copy is known as a crash dump. The crash dump information is saved in a set of files named unix.x and vmcore.x, where x is an integer that identifies the dump. You can configure the directory in which these files and can set other dump parameters using the dumpadm utility.

Parent topic: Logs and traces

Terms of use | Feedback
(C) Copyright IBM Corporation 2003, 2004. All Rights Reserved.