My HA instance seems broken after power outage

Configuration
1

My HA is hosted in Hyper-V on Windows 11. Since we had a power outage a few days ago HA has been unstable.

It works well right after the instance is booted up, but after an unknown amount of time (usually 1-2 hours), HA stops responding.

I load up the screen of the Hyper-V and see this error message:

ha > [ 3779.632997] CIFS: VFS: \\[HOSTSERVERNAME] has not responded in 180 seconds. Reconnecting...

If I restart the Hyper-V instance it comes back online until this message shows up again.

It indicates to me that there’s some issue with file access, maybe? \[HOSTSERVERNAME] refers to the name of the host machine, which is working as expected well though. So it seems to be related to the instance, hyper-v, the and some communications between them.

I checked the drive on which the Hyper-V instance disk is on and it contains no errors. I have another drive where HA backups are placed and no errors there either.

How would I go about debugging this?

Thanks!

PS. I have HA backups available, but I’d like to give it a go of fixing this before diving into creating a new instance and loading the backup… (although that might be the easiest?)

2

It should be added that the only Samba/Windows (CIFS) related I have in my setup (that I know of :sweat_smile:) is the setup of an external share to place said backups on. I’m going to try and set that up again, maybe something is off in that configuration now.

The above network location for backups is on a HDD, if that has any relevance.

The instance has 8GB ram available, currently using 1.2GB.

3
  • Restore a snapshot from before the power outage
  • backup HA, create a new VM, reinstall HA, restore backup
4

Yeah, I guess that’s the way for a clean no-frills approach.

I’m away from the office now, so I unfortunately can’t check if I’ve set up snapshots for the instance or may have neglected it. If they’re not available, restoring HA backups will be the way to go.

Thanks for the input :+1:t2:

5

I did this now and it worked surprisingly well - kudos to the HA team for an amazingly smooth restore process. All seems to be working again as desired, knock on wood.