Hello,
Hey, for future, um, not well informed, admins. If you hose your local attached storage used by an SR, you may need to drop back and punt. This was a development xcp-ng host with an XOA instance using the locally attached storage for backups.
So, no significant loss if I need to drop the SR and re-set up a VDI for xoa-backups.
After forcing a reboot during an xcp-ng upgrade, I failed to repair the disk. I will not argue foreign the host to hardware reboot was stupid. That failed miserably; the LVM partition would not fsck, even after leaving it for days. Nor from GUI could I delete the SR - it would just spin and do nothing - not even an error message in the GUI logs.
So after a bit of research hand find a helpful post by @olivierlambert logging into the xcp-ng instance and:
xe sr-forget uuid=<<UUID of screwed up storage>>
I then recreated the VDI for the XOA instance. Finally, I checked to ensure the new VDI was attached, created a new ext4 partition for the xoa-backups, verified it mounted, and restarted the backups.
Please ask if you find yourself stuck and need any clarification on any of this.
Regards,
Bill