We have a cluster pair (Active - Standby) of NX5500 controllers running on 184.108.40.206 firmware
They're supporting a mixture of AP6532 and AP7632 APs, totalling 104. The AP6532 APs stop us from upgrading the firmware, until they are replaced.
We monitor various parameters with snmp. We're seeing a lot of gaps with the snmp traffic. We suspect that this is because the primary controller is running with such a high cpu % and dropping snmp packets.
I would like to confirm if this is the case.
A service show top shows us this:
PID STATUS RSS PPID %CPU %MEM COMMAND
761 R 78248752 29.3 1.9 cfgd
We run two onboard captive portals, which I suspect may be causing the high CPU.
Guest (runs on all APs)
Staff (runs on only two APs in the Cafe)
The average CPU % has been increasing steadily over the last few months, but the Free RAM has stayed steady at 75.7% the entire time.
Is there anything we can do to try and reduce the average CPU % ?
Please review the Knowledgebase (KCS) article Christoph previously provided. Also, it is best practices to have two separate management policies. One for the NX controllers and one for the APs. Ensure that you are not polling the APs via the AP Management policy (disable SNMP), and just have SNMP enabled via the NX Management policy. Check your interval timers and ensure they are +30 minutes. Extensive SNMP polling (low interval timer) could be contributing to the issue.