On one of our R80.30 appliances we have an intermittent issue with the monitoring which we have traced down to certain OIDs from the .2620 tree which fail to respond in time. Interestingly enough, the issue does not appear to be with SNMP as such but rather with CPD as it fails a call to some libraries. Specifically, the cpd.elg file is filled with messages like the below:
[CPD 11415]@FW[8 Apr 16:45:20] Warning:cp_timed_blocker_handler: A handler [0xf67cd980] blocked for 10 seconds.
[CPD 11415]@FW[8 Apr 16:45:20] Warning:cp_timed_blocker_handler: Handler info: Library [/opt/CPshrd-R80.30/lib/libcpstatext.so], Function offset [0x13980].
[CPD 11415]@FW[8 Apr 16:45:35] Warning:cp_timed_blocker_handler: A handler [0xf12c2e90] blocked for 6 seconds.
[CPD 11415]@FW[8 Apr 16:45:35] Warning:cp_timed_blocker_handler: Handler info: Library [/opt/CPsuite-R80.30/fw1/lib/libfwstatagent.so], Function offset [0x7e90].
[8 Apr 16:45:35] CPStatGet: function excessive time. oid is: 1.3.6.1.4.1.2620.1.1.25.6
I have noticed that whenever such logs pop in the file, CPD hits 100% utilization on its CPU core. Some 95% of the logs are for issues with OIDs from 1.3.6.1.4.1.2620.1.1.25. subtree. Any ideas about the possible cause would be appreciated. 🙂
