Basic setup at Production site:
2x Dell R730 running Windows 2012 R2 Standard (Windows updates current as of June 2016). Cluster passed Validation tests, with the exception of some warnings due to iSCSI NICs not having a default gateway (those NICs can't ping the LAN interfaces)
3x Cluster Shared Volumes (one Witness, two data) with a single Generic Service role. Generic Service role has both data volumes as dependency. All CSVs are located on a Dell EqualLogic iSCSI SAN group; data volumes are replicated regularly to DR site.
For backups, we have scripted mounting recent replicas on the DR EqualLogic group to a server at the DR site. This setup has been working fine; however, the CSV replicas, when mounted to the DR system, have their filesystems marked as dirty. The filesystems on the Production CSVs are not marked dirty (verified by fsutil and Action Center). This happens occasionally as all replicas/snapshots are taken on the hardware side and the filesystems are not quiesced (which is OK for our purposes), but has been occurring on all replicas this past week.
Issue:
To ensure the CSVs were clean I opted to run a read-only chkdsk from the command line "chkdskX:" on each volume (no switches specified). When I did so, both CSVs were taken offline, which is contrary to my understanding that 2012 R2 CSVs should remain online for analysis and spotfix. On two other servers (not cluster systems), mounting a replica and running chkdsk, the volumes do stay online, so I'm a bit puzzled why the Cluster decided to take the volumes offline, and how to prevent this from happening in the future.
Do I need to specify the "/scan" switch or some other parameter? Does this need to be run through Action Center or Server Manager? Do I need to be running a File Server role in the Cluster?
I appreciate any help and thoughts on this!
Thanks!