14 minutes ago, Benz0187 said:

but guessing i´m doing it wrong...

With the array started type:

 

btrfs rescue zero-log /dev/mapper/sdh1

 

Then re-start the array and post new diags.

52 minutes ago, JorgeB said:

With the array started type:

 

btrfs rescue zero-log /dev/mapper/sdh1

 

Then re-start the array and post new diags.

Thnxz! that worked! Just have to ask, where did you find "sdh1"

Just rebooted and its the same problem again. Works with that command thoe...

 

oden-diagnostics-20240806-2153.zip

After that same command and array-restart it works again for now.
My dockers acting strange after this, some complain about configfiles and some wont start. Its something strange...

 

12 hours ago, Benz0187 said:

My dockers acting strange after this

 

Pool is going read-only, it has other problems, recommend backing up and reformatting the pool.

 

 

Was unable to load content to Plex from my server today. Checked the server and everything looked fine and then I caught a warning message about a Cache Drive Error. I stopped the array. I unmounted the noted cache with the error and restarted the array. Now I see the same "Unmountable: Unsupported or no file system" message next to my cache drive. Whats-more, when I stopped the array again to re-add the drive (with the initial error) from the drop down in the cache pool, it listed the drive as 0 b. I added it nonetheless, got the same Unmountable message. Stopped array again. Removed the (cache) drive with initial error. Still same (Unmountable) message but now the drive with the error no longer appears when I got to try and add a second cache drive to the pool.

 

I tried a few suggestions here, like btrfs rescue but those were more shots in the dark and lazy troubleshooting. I'm attaching the diagnostics to see if I can get some insight into a fix. Additionally, the array continued to experience issues when rebooting. It woul dbe stuck unmounting an array drive and I would have to manually unmount it. A quick terminal check showed a docker image holding up the process. When I check the docker status, it showed stopped. Then I would manually run umount /dev/loop2 (where dock img is located) and the reboot would work.

beedoops-diagnostics-20240907-1315.zip

Disable Docker and VM Manager in Settings until you have cache fixed. 

 

On mobile now so can't look at Diagnostics yet. 

Both have been disabled now. Attaching new diagnostic in case it's needed as I'll be mobile for a few hours.

beedoops-diagnostics-20240907-1459dockervm-off.zip

Post the output from:

 

btrfs fi show

 

Hi All,

 

I am having the same issue as most people here. I suspect the culprit has something to do with the docker "Immich" as I recently configured it AND with nvidia GPU. Everything was fine uploading images to Immich via external libraries (via my shares) until I added a library with a huge amount of images. The method I was using was add a library > scan for new items > search for items NOT in an album > finds a few hundred, but I scroll down so it loads more so I can add more at a time, otherwise it would take forever adding them. It was during this process it all the dockers stopped

 

I have 2 x 2tb nvme's in a pool which I am using for my dockers. When the issue occurred, I noticed both of them said "Unmountable: Unsupported or no file system"

 

I checked the logs and it looks like both nvme0n1 and nvme2n1 has some corrupted files.

Sep 19 14:21:47 Tower kernel: BTRFS info (device nvme0n1p1): bdev /dev/nvme0n1p1 errs: wr 0, rd 0, flush 0, corrupt 111, gen 0
Sep 19 14:21:47 Tower kernel: BTRFS info (device nvme0n1p1): bdev /dev/nvme2n1p1 errs: wr 0, rd 0, flush 0, corrupt 75, gen 0

Also noticed form After reading through this forum, I'd thought I'd try out some of the suggestions:

I ran the following:

btrfs rescue zero-log /dev/nvme0n1p1

Output:

Clearing log on /dev/nvme0n1p1, previous log_root 1952935706624, level 0

I did the same for the other nvme:

btrfs rescue zero-log /dev/nvme2n1p1

Output:

Clearing log on /dev/nvme2n1p1, previous log_root 0, level 0

 

This mounted my drives again, however, only a quarter of my dockers would seem to start but not actually run.
If I stop the array again, and start it back up, both drives revert back to "Unmountable: Unsupported or no file system"

I tried removing immich after it "starts", but i just get "Execution Error". Plex is using Nvidia gpu for transcoding, but it was never an issue and have been using it for years. I noticed my docker page seems to have a spasm every 2 seconds as well as thought the whole page is refreshing.

 

I haven't tried unmounting and remounting so to speak as I'm not sure what effect that would have. 

Have attached my logs. I hope someone can help!

 

tower-diagnostics-20240919-1703.zip

Edited September 19, 20241 yr by Daburgers

2 hours ago, Daburgers said:

Have attached my logs. I hope someone can help!

Since btrfs is detecting data corruption on both devices, first thing I would recommend is to run memtest.

 

Then run btrfs rescue zero-log again, and if it mounts, backup what you can from the pool and then re-format.

22 minutes ago, JorgeB said:

Since btrfs is detecting data corruption on both devices, first thing I would recommend is to run memtest.

 

Then run btrfs rescue zero-log again, and if it mounts, backup what you can from the pool and then re-format.

Thanks mate. I thought I'd kickoff the backup while I waited for a someone to reply and while the drives are currently able to mount, at least for now.

Will do a memtest as soon as its done and report back if anything is found. If not, I'll reformat.

Since I am just copying the contents of my appdata folder to one of my shares. Will it be as simple as copying and pasting it back in? appdata was all I had on those drives (plus some downloads which aren't important and will discard)

Edited September 19, 20241 yr by Daburgers

4 minutes ago, Daburgers said:

Will it be as simple as copying and pasting it back in?

Yes.

Have a similar issue. Cache is showing unmountable. I tried a few of the suggestions earlier in the thread. Any assistance would be greatly appreciated. tower-diagnostics-20241231-2123.zip

Edited January 1, 20251 yr by valhalla540

8 hours ago, valhalla540 said:

Have a similar issue.

Type:

 

btrfs rescue zero-log /dev/nvme0n1p1

 

Then restart the array and post new diags.

Got this message after running the command you provided. And the cache returned. Running some test, restarting plex, server, etc. Thank you for that. I tried the command earlier but I had stopped the array and didn't have the "p1" at the end. 

 

"Clearing log on /dev/nvme0n1p1, previous log_root 1528679710720, level 0"

 

Re-ran diagnostic

 

Thank you so much for your help. 

 

Edit: Unintended side effect I think is that all my dockers reverted back to stuff saved on the cache I think and is now inaccessible.

tower-diagnostics-20250101-1107.zip

Edited January 1, 20251 yr by valhalla540

Docker image is corrupt, but first run a correcting scrub for the pool and post the results.

I have the same problem. I formatted my cache drive several times but still the same problem. Attaching my diagnostics file if anyone is able to help. Thank you!

adelinetower-diagnostics-20250109-1633.zip

7 hours ago, Krum4o said:

I formatted my cache drive several times but still the same problem.

Device is busy and failed to format, reboot and try again.

12 hours ago, JorgeB said:

Device is busy and failed to format, reboot and try again.

Thank you very much, JorgeB! It worked. I really appreciate the help. 

Hi all,

 

today I encountered the same problem.

I tried everything from the recovery steps and also the btrfs rescue zero-log part but with no succes.

 

I get errors like

mount: /temp: can't read superblock on /dev/nvme0n1p1

 

and lots of checksum Errors whem attempting to recover the data via btrfs restore.

 

Any help or troubleshooting steps would be appreciated.

Thanks in advance

unraid-diagnostics-20250111-1243.zip

3 hours ago, Roam1988 said:

Any help or troubleshooting steps would be appreciated.

Thanks in advance

Jan 11 12:39:16 Unraid kernel: BTRFS error (device nvme0n1p1): bad tree block start, mirror 1 want 239517696 have 0
Jan 11 12:39:16 Unraid kernel: BTRFS error (device nvme0n1p1): bad tree block start, mirror 2 want 239517696 have 0

 

Tree block is empty, in both metadata copies, suggesting something was wiped that it shouldn't have been, if btrfs restore doesn't work, I'm afraid that there aren't many other options, you can try posting in the btrfs IRC channel, or using a file recovery app like UFS explorer.

Oh, that sounds realy bad.

Do you have any suggestions what could cause this behavior? 
The last thing yesterday was a sheduled shutdown and after booting today this occured. 
SMART seams fine and the drive isnt old either.

Should i replace the drive or do you think its more of a config/software issue?

5 minutes ago, Roam1988 said:

Do you have any suggestions what could cause this behavior?

Typically, it's an issue with the device firmware, where some data is missing/deleted that it shouldn't be.

Thank you for the fast replies. I am currently on vacation so IRC isnt realy an option. Do you probably know if the btrfs mailing List ist still working?