Manual trace and log file collection

Trace and log files contain information that help troubleshoot problems. If you cannot collect diagnostic data for the IBM® Db2® pureScale® Feature by using the db2support command, you can collect trace and log information manually.

If you can use the db2support command with the -install parameter, you do not need to use the information in this topic to collect diagnostic information manually. You can manually collect diagnostic information that is similar to what you can collect by using the db2support command with the -install parameter, but collecting information manually requires more effort.

IBM Support requires you to submit the following files if they are available. Create a single .zip file or .tar archive of all of these files.
  • Files from the /tmp directory:
    • Current installation or instance log and trace files in the /tmp directory on the installation-initiating host (IIH). To ensure that the trace and log files that you collect are relevant to the Db2 installation and configuration commands that returned errors, reissue any commands that failed with the trace option turned on (typically by including the -d or the -t parameter). For example, if you reissue the db2icrt command, include the -d command parameter to turn the trace on. The trace output file is named /tmp/db2icrt.trc.pid, where pid is a process identifier.
    • Other trace and log files in the /tmp directory:
      • The /tmp/db2trc.trc.* files
      • The files in the /tmp/db2log.pid directory on the IIH
      • The /tmp/ibm.db2.cluster.* files that were generated on the IIH during installation
      • The files in the /tmp/db2log directory on all remote hosts
  • Files from the $HOME/sqllib directory:
    • The $HOME/sqllib/db2nodes.cfg file
    • The $HOME/sqllib/db2dump/db2diag.log file
    If the $HOME/sqllib directory does not exist, collect the /tmp/db2nodes.cfg* and /tmp/db2diag.log* files.
  • The /etc/services file from each host.
  • For each node, the errors and trace logs for IBM Tivoli® System Automation for Multiplatforms (Tivoli SA MP) from the /var/ct/db2domainname directories and the error and trace logs from the /tmp/db2_cluster_manager_spooling/ trace spooling directory.1
  • Output of the following commands:
    • db2instance -list, issued as the instance owner
    • lssam > /tmp/lssam.out
    • On AIX®: 
         lslpp -l all | grep "sam." > /tmp/lslpp_sam.out
   lslpp -l all | grep gpfs > /tmp/lslpp_gpfs.out
   lslpp -l all | grep rsct > /tmp/lslpp_rsct.out
    • On Linux®: 
         rpm -qa | grep "sam." > /tmp/lslpp_sam.out
   rpm -qa | grep gpfs > /tmp/lslpp_gpfs.out
   rpm -qa | grep rsct > /tmp/lslpp_rsct.out
      • lsrsrc IBM.Equivalency > /tmp/lsrsrc_IBM.Equivalency.out
      • lsrsrc IBM.Application > /tmp/lsrsrc_IBM.Application.out
      • db2greg -dump, from all hosts
      • ps -fe | grep db2 and ipcs, from all hosts
      • id, issued as the instance owner and the fenced user on all hosts
      • uname -a, from all hosts
    • On AIX: 
         errpt -a, from all hosts
    • On Linux: 
         /var/log/messages, from all hosts
  • GPFS cluster file system logs from the /var/adm/ras/ directory.
  • Every file that is located within each node's /var/ct/iw/ directory and /var/ct/cfg/ directory.
  • A .zip file or .tar archive of all files in the instance owner's db2dump directory ($HOME/sqllib/db2dump).
1 For more information on Tivoli SA MP commands, see this link.