Although data is being collected and transferred to the console some of the metric groups are missing. Looking at the UDR Collection Manager (UCM) Status Reports some groups are reported in Terminated No Data' status and report the error:
Error creating semaphore - No more IDs available
This error message is also visible in the $BEST1_HOME/bgs/log/[hostname]-bgsagent_6767.log on the remote node.
NOTE: This document was originally published as Solution SLN000000226648.
Based upon the log messages the remote node has run out of semaphores so it can't allocate the metric groups.
If you have a compatible web browser on your console you should also be able to see these messages by opening the UDR Collection Manager (UCM) status reports via the following URL:
file:///usr/adm/best1_V.V.VV/local/manager/status/UCMStatus.htmlwhere V.V.VV is the product version, for example, '7.4.10'.
The following output would be useful to debug why this machine has run out of available semaphores:
The typical source of the problem is that the shared memory limits on that machine are somewhat low and either another shared memory consuming application has started on that machine and it is using up some of the IPC shared memory or some of the Perform shared memory segments have been orphaned and there isn't sufficient additional shared memory configured for a second allocation of segments.
The long term solution will likely be to increase the shared memory settings on the machine and reboot to get those settings applied. In the short term we might be able to manually clear some segments if Perform has orphaned them.
The recommended shared memory and semaphore settings for Perform data collection are documented in Resolution 15104702.