Hi,
We are using around 50 + cisco catalyst 2960X POE switches in our office Network for Wi-Fi & L2-functionality. All of the switches are configured for SNMP monitoring & sys logs(Observium & Zabbix).We had a Large vlan database of approx. 50+ vlan. SSH access is configured for remote login. All the switches are having uniform configuration apart from vlan database. Earlier we were getting continuous spikes in CPU utilization in all switches due to HULC LED, IP SNMP ,SNMP ENGINE & STP Processes. Last month I had updated the VTP(using VTP version2) configuration of all the switches from client mode to transparent mode to overcome the high cpu utilization due to STP, IP SNMP & SNMP ENGINE processes. I had manually configured only required vlans on switches as per requirement. It worked well for me and frequent cpu spikes had been stopped and cpu utilization had been dropped by 8-12% in 24P switches and 15-20% in 48 P switches. But since configuration changes I had observed some unusual behaviour from switches end. I had observed cpu spikes daily in around 20 switches at a fixed Interval of time in morning time. I am not aware whether earlier also same thing was happening or not(before configuration changes). It's happening since last 1 month. Upon checking the details i have found the main culprit i.e: ssh process (consuming 30-75% of cpu resource). Can anyone please tell me the root cause of this and how can i resolve it.
Thanks in Advance
