Back in the days of good V8.x there was a lovely bug that used to let BIBusTKServerMain runaway with loads of resources. I'm pretty sure there was a way to trace from the audit DB the Process ID that would correspond to the Process ID running on the server so we could kill that BIBusTKServerMain. Does this sound familiar ? can someone please confirm ? I have a number of rouge BIBusTKServerMain process that I would like to trace back to a user via the Audit logs, and kill the process. Is this still possible ??
We experienced the same issue just this week in our Development environment. We are running 11.0.9 but will be upgrading next month. The runaway process was creating tmp files and filling the disk. Stopping and starting services did NOT correct it, the server had to be rebooted. I think it may have been a bad report (I have several people authoring). I would also like to determine the cause.
If your into Linux please check the open files limitation.
1. Go to /etc/security
2. Open limits.conf file
By default values are below
s_cognos soft nproc 2047
s_cognos hard nproc 16384
s_cognos soft nofile 1024
s_cognos hard nofile 65536
s_cognos soft stack 10240
s_cognos hard stack 10240
Please modify the values like below
s_cognos soft nproc unlimited
s_cognos hard nproc unlimited
s_cognos soft nofile unlomited
s_cognos hard nofile 327680
s_cognos soft stack unlimited
s_cognos hard stack unlimited
s_cognos is nothing but your cognos service account.
Once you modify the values.
Reboot the server and start the cognos services.