On 31/03/2021 09:32, Sharon Kimble wrote: > > I'm hoping that you folks can help me with a problem that is now > happening reasonably regularly, actually twice. > > I have 2 data drives on my system /mnt/backa and /mnt/backb. Both are > 4tb drives, with backa being 2.74tb and backb 2.81tb. It might be helpful to see the relevant lines from your /etc/fstab. > > Backb is now regularly losing its mount point when I reboot, meaning > that only backa is mounted, and backb holds my restic backup. > > To regain access to backb I'm having to 'sudo e2fsck -y -b 32768 > /dev/sdc2' which rebuilds the mount point such that I can mount it on > reboot, as backb. Firstly, it's usually better to run the "fsck" frontend, which will determine which filesystem you've got and run the appropriate backend, but I can see that you're passing advanced parameters here, so jumping straight to e2fsck isn't that unusual. Secondly, you're specifying "-b 32768" which is telling fsck to use a secondary superblock. Why is this? Generally, the primary superblock should be "good enough" to repair a filesystem. The man page does state that the primary superblock *should* be updated after the fixes are complete, so this shouldn't be necessary. So, why are you having to use a secondary superblock? Do you know what's wrong with the primary one? > > So how do I stop it happening again please? And what is the cause of it > all? Should I physically unmount the drives before rebooting? The answers you seek should already be logged somewhere. Try the following: $ journalctl -b -u mnt-backb.mount # This will show output from attempts to mount /mnt/backb since the current bootup $ journalctl -b -g sdc2 # This will grep the journal for all messages containing "sdc2" since the current bootup > > Thanks > Sharon. >
Attachment:
OpenPGP_signature
Description: OpenPGP digital signature