CPU utilisation fluctuating on all nodes including publisher

[noc mcc]

Hi,

1.   We have a cluster having 12 nodes including pub and tftp. We are consolidating 4 different clusters into this one cluster.

All the required data of 4 clusters have been exported to this new cluster but phones are yet to be migrated to it.   But recently we have been facing an issue where cpu utilisation of all the nodes including publisher increases from 10% to 30% momentarily for approx. 30 sec and then come down. This causes continuous variation in cpu utilisation as seen in RTMT. CPU of publisher even rises to 50 % sometimes.  During this rise it is observed that cmoninit (and sometimes tomcat) process CPU is also rising.

2.  My concern is that when all the phones will be migrated to this new cluster will it effect the registration process of phones?

3.  we have given reset to all the nodes but still the issue persists.  No pending process is there in BAT.

4.  Any suggestions for the issues are welcome.

Thanks

[Andrew West]

Have you opened a TAC case? 

Have you pulled logs from RTMT or really perf mon values and tried to figure out whats happening?

Is Alert Central triggering when these spikes occur?

[Dennis Mink]

Can your spikes in Tomcat usage be tied in with it getting hammered through AXL?