Unraid v. 7.2.4

USER SHARES AND DOCKER CONTAINERS MISSING AFTER UNCLEAN SHUTDOWN.

I have appdata backup installed with backups available.

I have Unraid Config Guardian installed with backups available.

All of the backups that I have saved are from Unraid v. 7.2.3 as my server was recently upgraded.

Here's what happened:

I was in the process of data migration, moving media from a local windows machine to my Unraid server media share when my Unraid server stalled. Data migration stopped, and retry failed on the Windows machine. I then rebooted Windows and while Windows was rebooting I attempted to stop the Unraid array. I ended up with the Unraid logo cycling nonstop. I tried to refresh the page that I was viewing the WebUI from and still had the Unraid logo cycling.

I then pressed the power button briefly to do a soft shutdown of the server which failed. I then held the power button down, performing a hard shutdown.

After rebooting my Unraid server and accessing the WebUI, I received a notification that an Unclean Shutdown had occurred. All of my share drives are missing in addition to half of my docker containers, what appear to have been the containers running at the time of shutdown.

I have 4 drives, 1 parity, 2 HDD's, and 1 cache NVME, in addition to the USB Flash. All of the data stored on the array is viewable and accessible by browsing /mnt/disk1, /mnt/disk2, /mnt/cache, and /mnt/boot from the Main page on my server. So it appears that all of my data is still available. This is also how I am able to view the available backups for appdata backup and Unraid Config Guardian. All of my stored data, i.e. media and photos, is viewable using this method.

I have not done anything with my Unraid server other than trying to create a "TEST" sharable drive which failed.

What is the ideal way to restore my share drives and Docker containers prior to the unclean shutdown?

Please help!

I have attached diagnostics zip file for my server for reference.

I also have Syslog Server enabled to send logs remotely but don't know how to access them.

Yes, I am new to Unraid but a quick student. ;)

olivejuice-diagnostics-20260302-1129.zip

Edited March 2Mar 2 by Olivejuice

Looks like your NTFS cache is corrupt. Not sure Unraid has anyway to fix that. I guess you could put it in a Windows machine and try to fix it, but since it is nvme might be difficult.

Why are you using NTFS for cache anyway? I recommend reformatting it to one of the filesystems with more Linux and Unraid history and support - xfs, btrfs, zfs.

Then you can restore from your backups.

8 minutes ago, trurl said:

Looks like your NTFS cache is corrupt. Not sure Unraid has anyway to fix that. I guess you could put it in a Windows machine and try to fix it, but since it is nvme might be difficult.

Why are you using NTFS for cache anyway? I recommend reformatting it to one of the filesystems with more Linux and Unraid history and support - xfs, btrfs, zfs.

Then you can restore from your backups.

I really appreciate your prompt reply!

I am assuming that what you are suggesting is pulling the NVME out of the server, mounting it in a Windows machine, performing an error check and hope that Windows fixes the issue?

As for using NTFS, I didn't know that I should have changed this to xfs like the other drives, still learning...

Reformatting the NVME drive to xfs will wipe all of the data on it, this includes appdata and system. Should I copy these and re-write them to the NVME after reformatting to xfs?

Then restore using appdata backup or Unraid Config Gauardian?

I really appreciate your time and input!

I've never used Config Guardian. Do you have any VMs?

appdata backup should let you get your appdata back. That should be enough to get your dockers going again assuming you installed from the Apps page.

https://docs.unraid.net/unraid-os/troubleshooting/common-issues/docker-troubleshooting/

11 minutes ago, trurl said:

I've never used Config Guardian. Do you have any VMs?

appdata backup should let your get your appdata back. That should be enough to get your dockers going again assuming you installed from the Apps page.

https://docs.unraid.net/unraid-os/troubleshooting/common-issues/docker-troubleshooting/

Yes, I do have VM's but nothing that I couldn't re-installl, other than Home Assistant but that hasn't been completely setup either.

All of the dockers have been installed from the Apps page. I looked over the document that you shared for reinstalling the docker containers and one thing that Im noticing is that it mentions vdisk, I set Docker data-root to directory not vDisk. Does this change anything?

Back to my original post: will I have to recreate my share drives? If yes, will this delete all of the data stored on the drives?

1 hour ago, Olivejuice said:

I set Docker data-root to directory not vDisk. Does this change anything?

Not really, but I question why you would want to do that. Were you having problems filling docker.img?

1 hour ago, Olivejuice said:

Back to my original post: will I have to recreate my share drives? If yes, will this delete all of the data stored on the drives?

Your data is there, the corrupt cache has just broken user shares. After you get that fixed your user shares will all be back as they were, except possibly whatever will be missing from cache.

17 minutes ago, trurl said:

Not really, but I question why you would want to do that. Were you having problems filling docker.img?

I had seen to setup Docker data-root as "directory" vs vDisk in IBRACORP's video Zero to Hero #3 https://www.youtube.com/watch?v=t6Wb8ATE-lg&t=1239s

Was this wrong? Should I change this?

17 minutes ago, trurl said:

Your data is there, the corrupt cache has just broken user shares. After you get that fixed your user shares will all be back as they were, except possibly whatever will be missing from cache.

I was not able to run chkdsk on Windows because Windows does not detect the NVME cache drive with the Unraid appdata and system file folders on it. I didn't want to reformat the drive for fear of losing the data on it.

I then mounted the NVME cache back into the Unraid server and went into maintenance mode when starting the array. I went into the cache and performed a Check Filesystem Status from Cache Settings and that came back as "no file system corruption detected".

I then went into Self-Test and and performed a SMART short self-test and that came back "Completed without error"

I then performed a SMART extended self-test and that came back "Completed without error"

I then opened SMART error log and it shows this:

I created a cache backup with the appdata and system folder files and saved it to /mnt/disk1/cache_backup incase I need to reformat the NVME drive and then potentially change it to xfs format as it is currently NTFS.

What should I try now?.... I'm really not sure what to do, all I know is that I don't want to recreate everything from scratch.

So far all of the data is still accessible.

Edited March 3Mar 3 by Olivejuice

8 minutes ago, Olivejuice said:

setup Docker data-root as "directory" vs vDisk

Watch at least the first several minutes of this, it does a pretty good job of explaining why I don't recommend directory.

15 minutes ago, Olivejuice said:

So far all of the data is still accessible.

Your user shares are working now?

Just now, trurl said:

Watch at least the first several minutes of this, it does a pretty good job of explaining why I don't recommend directory.

I will watch this now

Just now, trurl said:

Your user shares are working now?

No, nothing is showing up and I am not able to create new shares. When I try to create new shares "test" and click ADD SHARE, Unraid processes the request but nothing shows up.

But then nothing shows up in the shares folder.

Your user shares are still broken due to the corruption on NTFS cache. Copy what you can and reformat.

Just now, trurl said:

Your user shares are still broken due to the corruption on NTFS cache. Copy what you can and reformat.

Ok, seems I'm making some progress.....

Like I mentioned, I had made a backup and saved it on /mnt/disk1/cache_backup. I then enabled Destructive Mode in Unassigned Devices. I then went into the Main page and removed the NVME from the pool. At this point there was a RED X next to the dive. Pressing the RED X cleared the drive allowing me to format it.

Once I formatted the drive I added it back to the pool and started the array. When I went back into my SHARES page all of the shared drives showed up, with the exception of my appdata share. Additionally the "Test" shares that I created earlier, were visible, which I then deleted.

I then enabled Docker in the Settings tab and the appdata share magically appeared in the SHARES page along with all of the others. I opened the appdata share and all of my app data was there. I then went to the PLUGINS page and went into Appdata Backup and restored the appdata for my templates and configs to a backup from a week ago. I went back into the Docker tab it only had Notifiarr in it which was running. I was able to open up the webui without an issue.

I then went into CA apps to begin reinstalling all of the containers, starting with RustDeskServer. I then went back into the DOCKER page and now magically Custom Folders was running and functioning like it was prior to all of this, with RustDeskServer running in its respective folder. Notifiarr was now missing. I went back into CA apps and going to Previously Installed, I installed Plex. I went back to the DOCKERS page and Plex was in its custom folder, running. I opened up the webui and everything was functioning as it should.

I will continue reinstalling all of the previously installed CA apps and update this thread as needed or if I run into any issues.

@trurl Thank you, Thank you!!!! You were a tremendous help and pointed me in the right direction, providing input along the way, to resolve my issue. You're an invaluable resource and I am thankful that you were here to help me. I will do my best to pay this forward through helping others when I can!

Thank you again!

Edited March 3Mar 3 by Olivejuice

@trurl Question: How were you able to determine that my NVME cache was corrupted from the diagnostics zip that I uploaded?

Also, should I change my Docker data-root to xfs vDisk? How can I do this without losing any data?

I learned that creating a vDisk limits the size that Docker will use which makes troubleshooting easier later if needed. I also learned that when Docker is set to run in a vDisk, it pulls only the missing layers that have not been pulled previously making it much faster and taking up less space because all of the layers are shared between the containers. Why would IBRACORP suggest using Docker data-root as directory instead of vDisk? He mentions in his video that it was because of the comments he received but doesn't explain why.

Thank you so much for your time!

Edited March 3Mar 3 by Olivejuice

Here's something that I ran into:

My custom network is missing and not selectable from the Network Type within the config of the container.

I also opened the terminal and typed "docker network ls" and it is missing.

I am assuming that I just recreate this and set all of the containers to my custom network? Why would I have lost my custom network? thoughts?

The only reason that Plex worked correctly is because the netwok was set to "host".

Edited March 3Mar 3 by Olivejuice

16 minutes ago, Olivejuice said:

just recreate this and set all of the containers to my custom network?

https://docs.unraid.net/unraid-os/troubleshooting/common-issues/docker-troubleshooting/#restoring-docker-custom-networks

1 hour ago, Olivejuice said:

How were you able to determine that my NVME cache was corrupted

Take a look in the logs folder of your zip.

Might be instructive to get new diagnostics and post them.

I recreated the custom network that I had before as it went missing during this whole process. The remaining containers that I installed were configured correctly, having used the custom network that I created. All of the previously install containers that missing the custom network, I changed in the config to my custom network and are all functioning correctly now.

I am doing a server reboot at the moment and will then post a new diagnostics zip.

Hopefully this will be my last question in this thread, so we'll see.

All of my shares are now set to the cache as the storage location. Obviously this is incorrect. I am under the assumption that appdata and system should be set to cache but what about all of the others. Previously I had set most of my shares to array and a couple of them to cache-array but I was wondering if there is a preferred structure for this?

36 minutes ago, trurl said:

Might be instructive to get new diagnostics and post them.

Here is the new diagnostics file:

olivejuice-diagnostics-20260303-1005.zip

7 minutes ago, Olivejuice said:

I recreated the custom network that I had before as it went missing during this whole process. The remaining containers that I installed were configured correctly, having used the custom network that I created. All of the previously install containers that missing the custom network, I changed in the config to my custom network and are all functioning correctly now.

I am doing a server reboot at the moment and will then post a new diagnostics zip.

Hopefully this will be my last question in this thread, so we'll see.

All of my shares are now set to the cache as the storage location. Obviously this is incorrect. I am under the assumption that appdata and system should be set to cache but what about all of the others. Previously I had set most of my shares to array and a couple of them to cache-array but I was wondering if there is a preferred structure for this?

I found this on primary and secondary storage but it doesn't state preferred configuation for storage:

https://docs.unraid.net/unraid-os/using-unraid-to/manage-storage/shares/#primary-and-secondary-storage

Edited March 3Mar 3 by Olivejuice

I had a lot of further advice, but I will save that for later discussion.

Your cache is still formatted NTFS. Why?

5 minutes ago, trurl said:

I had a lot of further advice, but I will save that for later discussion.

Your cache is still formatted NTFS. Why?

:( Because I wasn't sure if changing the structure, NTFS to XFS, would cause any issues with my current setup and considering that I was already having issues with my shares, I wanted to go with what was working previously.

Can I reformat the cache drive using the method I used previously, i.e. (stop array, remove device, clean disk, format) and format as xfs, then move the backup of my appdata and system back onto the newly formatted xfs cache and not have any issues?

3 minutes ago, Olivejuice said:

:( Because I wasn't sure if changing the structure, NTFS to XFS, would cause any issues with my current setup and considering that I was already having issues with my shares, I wanted to go with what was working previously.

Can I reformat the cache drive using the method I used previously, i.e. (stop array, remove device, clean disk, format) and format as xfs, then move the backup of my appdata and system back onto the newly formatted xfs cache and not have any issues?

Attempting to change this now.... keep you posted.

52 minutes ago, Olivejuice said:

:( Because I wasn't sure if changing the structure, NTFS to XFS, would cause any issues with my current setup and considering that I was already having issues with my shares, I wanted to go with what was working previously.

Can I reformat the cache drive using the method I used previously, i.e. (stop array, remove device, clean disk, format) and format as xfs, then move the backup of my appdata and system back onto the newly formatted xfs cache and not have any issues?

Cache is reformatted to XFS.

Here is the new diagnostics zip:

olivejuice-diagnostics-20260303-1140.zip

Edited March 3Mar 3 by Olivejuice

17 hours ago, Olivejuice said:

I had seen to setup Docker data-root as "directory" vs vDisk in IBRACORP's video Zero to Hero #3 https://www.youtube.com/watch?v=t6Wb8ATE-lg&t=1239s

Was this wrong? Should I change this?

This was question that I posted earlier in the thread but didn't receive a definitive answer.

Should I change the Docker data-root to vDisk instead of directory? How do I do this properly?