I have a scenario I haven’t prepared for, and need to think through it. We have a VMWare environment whose storage is on multiple NAS devices. We use a TrueNAS Scale (Bluefin) box for bulk storage of archived videos. This is mostly a windows environment.
We have a single ZVol which is passed to the VMWare cluster, and drives happily exist. I’ve been taking ZFS snapshots every hour, they persist for a few weeks.
We need to get a file off a VM from a snapshot. I’m realizing I’ve never gone through it, and it didn’t go how I planned. I know I can’t see the iscsi objects in the filesystem like you would a dataset for NFS sharing, so we’re not going in and directly touching VMDK’s or anything. I can clone the snapshot to a new dataset, but then when I go to create a new iSCSI connection for it (figuring I’d have to connect that to VMWare, attach the VMDK to the server as an extra drive, copy the file out, disconnect the drive, disconnect the iSCSI storage, promote the original dataset, destroy the snapshot) I don’t have the option to add an existing dataset to iSCSI sharing. My research is showing people saying that can’t be done.
Is there a method to do this (other than rolling the whole dataset back)? These are huge disks (5-10TB) for video archiving of classroom recordings so I don’t have enough spare space to work with to clone the whole dataset.