Post Fix Pack Blues

[NYD3030]

Cognosoids,

I'm currently applying Fix Pack 12 to a Cognos 10.2.1 instance and running into some issues following installation. For some background, this is two dispatchers running on a single VM, with a separate VM for the gateway. The fix is being applied to base Cognos 10.2.1 with interim fixes 12 and 13 applied by a previous resource some years ago.

1.   The query services is continually crashing and re-launching, at a rate of about once every two minutes, and the log seems to indicate a memory issue. There are no DQM reports in this environment. The error states: DQServer: shutdown due to the OutOfMemoryError
2.   While it's not manifesting itself obviously in the environment, the catalina log is filling up with 'failed to get a RMI stub' errors.

Overall there are no issues in cognos connection or running reports, but users are reporting that post-patch performance is poor.

I'm tried some things but nothing has helped - any insight would be appreciated.

[NYD3030]

I left out a crucial piece of information - these issues only occur when both dispatchers are running. There are no errors with a single dispatcher.

[MFGF]

Cognosoids,

I'm currently applying Fix Pack 12 to a Cognos 10.2.1 instance and running into some issues following installation. For some background, this is two dispatchers running on a single VM, with a separate VM for the gateway. The fix is being applied to base Cognos 10.2.1 with interim fixes 12 and 13 applied by a previous resource some years ago.

1.   The query services is continually crashing and re-launching, at a rate of about once every two minutes, and the log seems to indicate a memory issue. There are no DQM reports in this environment. The error states: DQServer: shutdown due to the OutOfMemoryError
2.   While it's not manifesting itself obviously in the environment, the catalina log is filling up with 'failed to get a RMI stub' errors.

Overall there are no issues in cognos connection or running reports, but users are reporting that post-patch performance is poor.

I'm tried some things but nothing has helped - any insight would be appreciated.

Hi,

What are the startup / maximum memory values set to be for the Query Service on each of the dispatchers? You can see this in the properties of each of the Query Services in the Admin console. Is it possible the VM doesn't have enough memory allocated to support both query services running?

If you have no DQM reports at all, you can easily disable the query service for each dispatcher - go into the Cognos Configuration utility for each and select the 'IBM Cognos services' option on the left. For the 'Query Service enabled' entry, set the value to False, then re-start the Cognos service for the dispatcher. That will save you some memory on the VM.

Cheers!

MF.

[NYD3030]

Thanks for the quick response! The startup/max memory settings were defaulted to 1028. I increased that to 4096 and bounced the dispatchers to no avail. The VM in question has 32GB of memory so I don't think it's a resource constraint. I also tried upping the memory for Tomcat on a lark, no improvement. As a side note, any idea what that should be set to? I can't find clear info on why I'd want to raise it or what it does exactly.

I'm starting to think the out of memory message is a red herring but I'm not sure. There are also 'connection refused' errors popping up. I have until 3:00 eastern today to get this figured out or else I'll have a displeased client...