Hello everyone, 

 

today I got a notice from my own uptime-kuma, that my nextcloud is down. 

After checking my mariadb, I noticed, that I can't connect to the DB via:

mariadb - uroot -p

 

While looking into the problem, I noticed that some appdata folders on my chache-drive are gone. But only for stopped docker containers.

My Cache disks are still "healthy" and the utilization is at 11%.

It's really weird and I'm kind of scared. Luckily I have backups on my ZFS-Pool. But before I do anything, I wanted to ask for help.

 

Here are my diagnostics:

 

 

unraid-diagnostics-20250304-1126.zip

I think I figured it out.

It looks like these errors are caused by a docker problem on unraid.

The docker network hostname resolution is not working. 

I configured my NPM to use the IP instead of the hostname for my other containers and everything works. 

Since the /mnt/cache/appdata/ directory is the main docker directory, the folders for the stopped containers could be deleted by docker. 

Now, I just have to figure out a solution.

Disable and enable docker didn't help

Looks like your docker uses /mnt/cache but for some reason your share is setup to use the cache instead of being cache only. So every file thats not in use is being moved by the mover to the disk and thus those files are gone for everything expecting it to be at /mnt/cache

Edited March 4, 20251 yr by Mainfrezzer

I also tried to rollback to 7.0.0.

DId not help either

31 minutes ago, Mainfrezzer said:

Looks like your docker uses /mnt/cache but for some reason your share is setup to use the cache instead of being cache only. So every file thats not in use is being moved by the mover to the disk and thus those files are gone for everything expecting it to be at /mnt/cache

Very good hint. I had to uninstall Mover Tuning, since it has been deprecated. 

So, @Mainfrezzerwas right. 
In the past my Mover copied all files from my cache appdata to my array as a backup. 
After uninstalling Mover Tuning, the Mover actually moved the files. 

Appdata of Containers that where stopped/unused were completly moved and all running containers crashed, because the Mover moved files that the containers used not so often. 
Luckily I had backup from the Backup / Restore Appdata Plugin. 

After restoring all files into a new folders, changing all containers to use those now folders, everything works again. 
My data is not gone, nextcloud and dbs are running.
Perfect. 

Man, that was scary.