ajburnet

Thanks folks - looks like an xfs_repair -L has made the file system readable again - haven't yet checked the data but hopefully all good.

Thanks - will try the xfs repair command on the device. If that works, how would I go about re-building the drive? Thanks again for the pointers. Amazing how much of this I have forgotten having not touched it for about 5 years!!

Ran it against the physical drive, e.g. dev/sdn. THanks.

Hi folks, I've had a look through the forum and FAQ and couldn't see anything obvious on how to proceed - its been a while since I've had an issue with my unraid setup. It looks like one of my drives has a filesystem issue - its showing as a non mountable filesystem. I've run an xfs repair but that ended reporting no superblock could be found. So not sure how to recover given I can't mount it and recover any data. If I format, will unraid just rebuild it from parity? Probably a dumb question but any advice on how to proceed would be gratefully received! Thanks, Alex

That fid it, thanks!

My folder containing my VM images etc was deleted last week and I'm now unable to start the kvm - if I look at the settings page, in the VM Settings section the status is shown as stopped. Any ideas on how to restart it (server restart has no effect). I can see in the log the following entries which look relevant! Oct 7 14:18:44 Labrador root: /mnt/cache/vms/ is not a file Thanks! Alex.

That would explain it......! Thanks, Alex.

Yeah, very strange behaviour. If it was some kind of symlink then deleting it in /user should result in the data being deleted on a seperate drive which is the behaviour I find strange. Damage is done anyway, docker.img, all appdata and vms now gone. Hopefully restoring from the backup works. But trying to stop the behaviour from reoccuring! Thanks!

Thank you for your input btw. What I don't get is the folders never used to exist in /user - i.e. they weren't a share (on purpose so the disks could spin down etc.). Now they are, and deleting them has completely wiped out the folder on the cache drive. In this case, the only share that it shows is the /system one for Vms, and its creating a folder in /user even though the share is set to only use cache and exclude all other disks. Does that make sense?

Seems like if I delete a folder that exists on /mnt/user/ it deletes the corresponding folder (if it exists) in /mnt/cache. Shh'd in as root so not a permissions issue, i.e. can't see the files. Eh?

Docker containers not showing up in the Docker UI in unraid now, but each container config path was mapped to a corresponding folder in /mnt/cache/appdata. All but two of those folders have now disappeared (e.g. I had Plex, CouchPotato, Sonarr, ZoneMinder - only Plex and Couchpotato folder existing in the cache appdata folder now). I had made no changes recently other than a reboot and Plex container update. Just rebooted again out of interest, and its deleted the entire appdata folder on the cache drive. This is really bizarre!

I can see it's recreating the docker.img folder and appdata folder on reboot in the /user folder, i.e. not on the cache drive. That's with the default docker storage location as: /mnt/cache/docker.img Odd...

Thanks - screenshot of the appdata share settings attached. Just rebooted and my containers have all disappeared too - I think it may have reverted to storing the docker.img off the cache drive as well....

Hi folks, I've always had my docker app config location as my cache drive (and the appdata share shows as using the cache drive only). Its never been an issue and data has always been there as expected. But something strange has happened recently - its saved everything (and duplicated in some cases) to an array based appdata folder. I (stupidly) deleted that yesterday thinking that was old info. Suffice to say, I totally deleted everything...which was nice. Anyhoo - I see the docker settings did not have the cache drive as default so I'm assuming unraid overrode the per app config file specified location on reboot at some point - does that sound right / expected behaviour? Thanks, Alex.

Hi - yes I'm running it using host networking. Have managed to SSH into the host and have somehow now claimed the server as its visible, thanks for your help!