Hello friends,

I'm facing a serious issue.
I upgraded my Unraid to 6.12.6 a few days ago and everything was working properly until now. Last thing I did with my system was update the mariadb docker app and then I restarted my system. I didn't add any new drives.

After I restarted my system I got an error message saying the cache pool is: "Unmountable: Unsupported or no file system".

From reading other posts on this matter, I added my diagnostics file.

Can someone please explain what I can do to fix this? 🙏🙏

tower-diagnostics-20240111-1549.zip

Edited January 11, 20242 yr by mjeshurun

If the log tree is the only issue this should help:

 

btrfs rescue zero-log /dev/sde1

 

Then restart the array.

Thank you very much,@JorgeBfor the quick response to my post.

using the command you included does bring back the cache pool back, but it only stays detectable till the next array restart.
If I restart the array another time the cache pool shows the same error message again.

Also, after I use the command you mention and the cache pool is detected, I'm seeing a docker service failed to start error message.

31 minutes ago, mjeshurun said:

If I restart the array another time the cache pool shows the same error message again.

In that case suggest backing up the pool and re-formatting, also recreate the docker image.

Can you please explain how best to backup the cache pool? 
And how to recreate the docker image?

Just copy everything somewhere else, like the array, you can use any tool like Dynamix File Manager or Midnight Commander for example.

 

https://docs.unraid.net/unraid-os/manual/docker-management/#re-create-the-docker-image-file

OK, thanks 🙏

Sorry for the noob questions, but after backing up the cache data, how do I re-format the cache drive pool?
After formatting do I just copy back all the data to the cache pool from the backup?

Edited January 11, 20242 yr by mjeshurun

When the pool is unmountable just click format next to array stop button, then move the data back.

Thank you, very much.

For some unknown reason to me, today the cache drives are no longer showing the "Unmountable: Unsupported or no file system" message.

I still got the "docker service failed to load" message. I renamed the old docker.img file to docker.tmp and restarted docker, and now docker loads properly, though with no containers.

Is there an easy way to bring back all my containers (without needing to manually reinstall each container using the template)?

Should I still format the cache drive pool even if it's mounting properly now?

You can restore them from CA:

https://docs.unraid.net/unraid-os/manual/docker-management/#re-create-the-docker-image-file

Thank you for the explanation 🙏

After renaming the old docker.img file to docker.tmp I managed to start the docker service and re-install the containers.

However, I found a few more problems that don't work as they did before the cache drives "unmountable" issue.

1. I have a few docker services I configured to work with NginxProxyManager. Although I was abale to reach these services in the past, they now fail to load with my external domain address. I can reach these containers' webui using my internal server ip address, but not with my external domain address.

2. if I set NginxProxyManager to use a custom network type, it opens the webui login page, but I can't log in. I can only log in if I set NginxProxyManager network type to Host. Can that be the reason I can't reach my containers using the external domain address?

3. VM's are also not working properly, and I'm seeing this error message when trying to go to the VMs tab: "Libvirt Service failed to start". Do I also need to delete/rename the libvirt.img file?

 

For the container issues suggest you ask for help in their respective support thread:

 

 

13 minutes ago, mjeshurun said:

3. VM's are also not working properly, and I'm seeing this error message when trying to go to the VMs tab: "Libvirt Service failed to start". Do I also need to delete/rename the libvirt.img file?

Reboot and if the service still doesn't start post new diags, note that if you delete libvirt.img all VMs will disappear (but not the vdisks)

Here is the new diagnostics file. Hope it helps shed some light on the issues I'm still having.

 

tower-diagnostics-20240112-1828.zip

Libvirt.img is corrupt, if you have a backup restore it, if not create a new image, but will then need to recreate the VMs, you can create then using the same settings and point to the existing vdisks, it should work for most.

Hi @JorgeB, I hope you don't mind if I trouble you with this issue a bit longer.
It seems I still have corruptions with my cache drives pool.

I've been having more unexpected issues with my system and after reading the forum I ran a filesystem check of the drives.

These are the results I got:

[1/7] checking root items
[2/7] checking extents
data extent[11303731200, 12288] bytenr mimsmatch, extent item bytenr 11303731200 file item bytenr 0
data extent[11303731200, 12288] referencer count mismatch (root 9351338399421693957 owner 16777096 offset 0) wanted 1 have 0
data extent[11303731200, 12288] referencer count mismatch (root 5 owner 9630909 offset 0) wanted 0 have 1
backpointer mismatch on [11303731200 12288]
ERROR: errors found in extent allocation tree or chunk allocation
[3/7] checking free space tree
free space info recorded 1385 extents, counted 1397
There are still entries left in the space cache
cache appears valid but isn't 965383028736
[4/7] checking fs roots
[5/7] checking only csums items (without verifying data)
[6/7] checking root refs
[7/7] checking quota groups skipped (not enabled on this FS)
Opening filesystem to check...
Checking filesystem on /dev/sde1
UUID: e0192f98-c083-4849-b444-e7c7a122f3e4
found 1210729578496 bytes used, error(s) found
total csum bytes: 429124228
total tree bytes: 1053720576
total fs tree bytes: 380370944
total extent tree bytes: 135462912
btree space waste bytes: 192168905
file data blocks allocated: 3048914657280
 referenced 1179449688064

 

Are these issues that can be fixed or do I need to re-format the cache pool, as you suggested initially?

13 hours ago, mjeshurun said:

do I need to re-format the cache pool, as you suggested initially?

I would recommend this.

OK, thanks. I will re-format the drives.

Any suggestions about drive configurations and format types?

I opened another thread concerning drive best practices, and I could use help to configure the system in a better way.

 

1. which format type is better for SSD's: btrfs or zfs?

 

Until now I ran my system with a btrfs cache pool consisting of two SSDs, without parity and without the use of an array drive for the mover function. Instead of using array and parity drives I backed up my data to an unassigned drive using luckybackup.

 

2. which practice is better, using a cache pool alongside an unassigned drive for backups, without the use of array and parity drives, or using 1 SSD for the cache drive, 1 SSD for an array drive, and a third unassigned drive for backups?

Both are good options but I would recommend zfs since it's better at handling dropped devices, still needs to be monitored.

 

That depends mostly on your use case/preference.