You may get spurious alarm messages that have the following text:
cdm:InternalAlarm: Unable to get Memory data (temporarily out of resources)
cdm:InternalAlarm: Unable to get CPU data (error)
cdm:InternalAlarm: Unable to get Memory data (error)
On AIX machines the commands CDM will run will timeout before completing. Changing the timeout will resolve the issue.
On some systems incorrect NTP settings can cause this issue
On Windows Systems the performance counters may have been corrupted
On AIX Systems:
If you are receiving CPU alarms - add the following in raw configure:
<cpu> section of cdm
add a new key if it does not exist with the following value
sliced_cpu_interval = 30
If you are receiving memory alarms - add the following in raw configure:
<memory> section of cdm
sliced_memory_interval = 30
On Windows systems:
Work with the system administrator to rebuild the performance counters. CDM probe uses these Windows components to collect the metrics.
If needed, go through the below link for “how to manually rebuild the performance counter library values “.
On both systems:
Check that NTP settings are correct and accurate. A few seconds of drift may cause this issue.