<up_the_irons>
maintenance on a different server last night didn't go as well... there was a left over linux_raid_member superblock on two NVMe drives that belong to ZFS pools. mdadm caught the drives on a scan during boot, and tried to put them into an array, holding them back from 'zfs import' ; this never happened before, presumably by coincidence (ZFS grabbed the drives first). So, the server came up
<up_the_irons>
without any customer volumes, or even /home. fml
<up_the_irons>
thankfully, was able to resolve it within the original maintenance window timeframe and got everything back up