I am running a standard file server cluster with two 2012R2 nodes and three 60disk high density SAS enclosures, I have 48 7200RPM disks in a non-Tiered 8 column double mirror pool. The rest of the enclosure is pooled for Tiered storage. Thus far I have carved out about 49 Volumes of various sizes (mostly 1TB using mount points). Things were fine until I added the last twenty volumes. I patched and rebooted the cluster during a maintenance cycle and that is when problems started to occur. Post reboot more than half the volumes went into a failed state. After 12 hours on the phone with MS was determined to be the cluster can't agree which node should own the volume.
I am still getting volumes that will drop into a failed state which, I have managed to get all but five volumes back up and every fail over is a tenuous event that doesn't guarantee all the drives will come back. The answer thus far is to move the disks into a separate role and fail the disks back and forth between the nodes until eventually the disk comes back online. This "recovery"process locks the systems up, and if the volume of disks is to high it has caused the cluster to hang and fail over unexpectedly which usually generates more failed drives.
This cannot be the only way to get this back. Has someone out there got any miraculous ideas? We are at the point of abandoning the solution and continuing the migration on a traditional SAN solution. Any help or ideas would be appreciated.