Dear All,
Since a long time now we have a problem in our 2012 R2 cluster.
Every hour we have a freeze of the whole cluster (6 VLANS with different environments) for 8-10 minutes.
The time is every hour around the same time. The monitoring tells us there is a high i/o wait, but there is no load that does explain the i/o wait. After the 8-10 minutes the whole cluster works again. This behaviour does only happen during production time.
At night when the clients are not working the behaviour is not there.
We updated all hardware and software with the latest versions. Also we increased the Queue depth and decreased theTimeOutValue (/kb/2806730/) We had an improvement during a small period of time, after we did this maintenance.
But the problem came back. It does not matter what vlan or CSV, all the Cluster is impacted.
We are almost sure it has nothing to do with the storage or with the VM´s, but with something in between.
But for now we are running out of ideas. Anybody ?