SM 9.34 Performance isues

We have recently performed Service Manager Upgrade from 9.31 to 9.34 (last month) and the application performance is very slow when compared to previous version. Major issue is users working on Quotes, as these are opening very slowly (say, 2-5 minutes to view or update Quote). We are also seeing performance issues in Incidents and Interactions but these are much better when compared to Quotes. Please let me know what changes to be done to increase the performance.

  • Hi,

    The best approach is to open a debugnode (e.g. sm -httpPort:<available port number> -debugnode:1 -RTM:3 -debugdbquery:999 -log:../logs/trace.log), login to Sm via the debugnode, reproduce the issue and upload the log file(s) for analysis. Please create separate log files for each issue e.g. 1 log file should contain opening Quotes, another log file will contain opening Incidents etc.


  • Hi,


    I think we need to start with the standard support questions to get a good grip on the issue:

    - The upgrade to SM 9.34: Was it binary only, or also with applications?

    - Was anything else upgraded at the same time - i.e upgrading Oracle to release 12


    Then typical analysis steps:

    - Inspect sm.log file for "RTE A" messages.

    - Inspect sm.log for "sqllimit" messages (as sqllimit parameter is set to 30 seconds by default very slow SQL statements will appear in the log)

    - Trace with debugdbquery:999, rtm:3. You can start a separate -debugnode servlet to limit trace to your session only. You can also use dynamic debugging feature to activate the trace for a single session (status monitor, "s" command for the session to trace, then click "Send Debug Msg" button).

    - If this trace is too heavy for production, you may want to trace with debugdbquery:5, rtm:2. This will not cause a lot of load will producing valuable information about the issue.

    These steps are for query and application performance issues - which are the ones we see very frequently. It also will help to identify issues like too large global list records, and so one. Web tier or web client performance issue will not be identifiable by these - you can look for these if query and application causes would have been ruled out.

    Best regards,

    Armin Franke