I've got an interesting problem that I think may be related to DNS. I have a 2 node OES2SP3 cluster (VMWare), pretty much identical (DNS,DHCP,File Services). Both have a 200GB cluster volume that normally resides on the node. SRV1 has CVOL1 and SRV2 has CVOL2. These servers/volumes servicing roughly equal loads across a 2000 node network. The only major difference that I can think of between the two servers is that SRV1 is the DNS designated master. Here's the problem. During the day when people are using the system, IF CVOL1 is resident on SRV1, CPU utilization peaks and usually stays peaked (top shows ndsd high util). So much so that after a short time my replicas can't seem to stay in sync. SRV1 is not the master, but does hold a read/write copy. If I migrate CVOL1 to SRV2 there are still occasional spikes in CPU on SRV1, but no problems arise from this and there seems to be no ill effects from having CVOL1 active on SRV2. The master replica is on a 3rd OES2SP3 that is NOT part of the cluster.

Is it possible that DNS is causing enough traffic to cause this problem? If so, is there anything I can do (settings, parameters, etc.) to mitigate? Thanks in advance.