I wanted to give everyone a heads up on an issue we spent weeks on trying to fix on Failover cluster.
For weeks we were getting these errors on Hyper-v. We worked with storage vendor (Dell) and Microsoft. Our shiny new just built 2016 Hyper-v cluster would randomly start having CSV issues. It would take all of the VM's offline. Random VM restarts. Only way to fix it was restart the entire cluster. The whole system seemed very unstable. Sometimes we would have to hold the power button down on the servers to restart them because they were so hosed up. Sometimes it would run a couple of weeks and have zero problems then have a melt down.
We would get these errors in failover cluster.
Cluster Shared Volume has entered a paused state because of 'STATUS_CONNECTION_DISCONNECTED(c000020c)'. All I/O will temporarily be queued until a path to the volume is reestablished.
We worked with Dell and Microsoft and no resolution. Then finally it donned on me that we didn't have these problems on our DR cluster until we made changes to the DR cluster to bring it up to Production cluster standards. One of the changes made on Prod was disabling and removing SMBv1 protocol. This was done on DR cluster and immediately we started having the same cluster issues as the production site. We didn't have before. A quick check of powershell get-smbserverconfiguration resulted in EnableSMB1Protocol being set to True. We found this odd since we removed the protocol from the servers. Well it appears Microsoft will allow you to remove SMBv1 but it will remain active in the OS until you run Set-SmbServerConfiguration -EnableSMB1Protocol $false. A reboot of the cluster and no more stability problems. It must be a windows bug.
For those of you that have unstable Windows Server 2016 Hyper-v Cluster check SMBv1.