I have a cluster in a data center that is doing power maintenance which has the possibility that all power will go out. The data center is performing the work midday over the course of a couple days so we cannot shut down preventatively.
My question is: Do the nodes do any sort of checkdisk once power returns, or does it just flush the NVRAM to disk and carry on (assuming the cluster comes up fine and there is quorum)?
Is there anything else I should expect or plan for when the cluster comes back online?
One of the core tenets of ONTAP and WAFL is guaranteed consistency on disk. No chkdsk or fsck on boot is required.
if you have FC or iSCSI luns mounted on hosts, and those hosts have not flushed their filesystems, those filesystems could require chkdsk or fsck from those hosts for those LUNs, since consistency of the filesystems in those LUNS is the responsibility of the hosts in these cases, not NetApp.