Hi to all,

as title, i have strange cpu spike every hour on exchange 2013 servers, here is the detail:

3x exchange 2013 (CU9), MBX+CAS, DAG (3rd node with 24h delay).

Clients connects only to node1 and node2.

Here is the graph from SCOM related to node1 and node2, cpu usage for "nonexchange (total)", node1 has all datastore mounted (5).

Some days ago the problem was related only to node1, i moved all datastore to node2 and for some days the cpu was ok on node2, but then, BAM, cpu spike problem also on node2.

Node3, holding delay dag (24h) and no client connection don't have spike problem.

"Exchange (total)" counter don't have spike:

Based on counter explanation, seems that cpu spike is NOT related to exchange process. But i'm not confident on that.

Nothing happen on servers, no warning, no critical alert. OS is Server 2012R2 clean, with only SCOM 2012R2 agent on it.

Someone could point me in the right direction in finding what's happen and to find what's the process that generate spikes?

Thanks!

Have you got hyperthreading enabled on CPUs? - I would turn that off and also check if any of the IIS (W3WP) processes are using up the CPU. The best way to check what is using cpu is to download processmonitor tool to verify what is using the CPU. https://technet.microsoft.com/en-us/library/bb896645.aspx HTH Abhi