[SOLVED] BTRFS error with new System / M.2 SSD

[Nxt97]

Hey everyone.

I swapped from Server Hardware to "more efficient" hardware some month ago. Since then every 1-3 days my logs are written full of errors and I need to restart Unraid. I don't see any problems other then the full log.

Of course I checked the log and it seems like there is some problem with my M.2 Drive or maybe with my Mainboard/M.2. I remember I wasn't sure when I ordered the Mainboard if it is able to use the full Bandwidth of the M.2.

 

Repeating Error: 

Jun 20 00:30:54 Zeus kernel: print_req_error: I/O error, dev loop2, sector 4975992
Jun 20 00:30:54 Zeus kernel: BTRFS error (device loop2): bdev /dev/loop2 errs: wr 5414, rd 0, flush 1, corrupt 0, gen 0
Jun 20 00:30:54 Zeus kernel: loop: Write error at byte offset 2773131264, length 4096.

Mainboard: ASRock Mainboard X470D4U2-2T Sockel AM4 mATX
SSD: Some WD M.2 that came with my Laptop.. But SMART stats seem to look ok. I want to upgrade to another one anyway soon and put this one back in my Laptop.

Also is there an unassigned SSD for VM's but I think this should not be a problem (It's the old cache drive)
CPU: Ryzen 7 3700X
RAM: 2x 16GB (1x 16384MB) Crucial CT16G4WFD8266 DDR4-2666 ECC DIMM CL19 Single

Maybe someone can help me, its sad that I can't rely on my Server at the moment.
Thanks in advance.

zeus-diagnostics-20200620-0016.zip

Edited June 20, 2020 by Nxt97
changed ram count

[JorgeB]

There's filesystem corruption on cache, best bet is to backup, re-format and restore cache data.

[Nxt97]

On 6/20/2020 at 8:28 AM, johnnie.black said:

There's filesystem corruption on cache, best bet is to backup, re-format and restore cache data.

Thanks everything is looking good for now.

 

If somone else got the problem, this is what I did:

 

 

• Disable the Docker under Settings->Docker->EnableDocker
• Disable the VM Manager under Settings->VM Manager->EnableVMs
• Backup Cache Drive:

/usr/bin/rsync -avrtH --delete /mnt/cache/ /mnt/(SOMEDISK)/cache_backup >> /var/log/cache_backup.log

• Format Disk:
  • Stop Array
  • Under Main -> Cache
  • Change "File system type:" to something different
  • Start array
  • Under the stop array button you should find a format button -> format disk
  • Stop array
  • Change "File system type:" back to the old value
  • Start array
  • Format again
  • Done
• Restore Cache Disk:

/usr/bin/rsync -avrtH --delete /mnt/(SOMEDISK)/cache_backup /mnt/cache/ >> /var/log/cache_backup2.log

DONE

Edited June 22, 2020 by Nxt97
changed code highlighting

[JorgeB]

Thanks for write up, and if you don't mind going to mark this as solved.

[Ymetro]

Thanks for this guide @Nxt97!

I had the same problem with 2 Sata SSD's as a BTRFS Cache Pool. 

For Unraid 6.8.3 I recommend the Unassigned Devices plugin to remove the old formatted partitions on the cache disk(s). 

So you:

• Do al the 3 steps Nx97's post mentions (disable VM Manager and Docker and start the rsync backup command)
• stop the array,
• deselect the SSD'(s) under Main -> Cache and it/they appear in the Unassigned Devices list below.
• Click on the plus sign [+] on the disk(s) in that UD list and it wil fold out showing the partition in it, which you can
• remove by pressing the Remove button. 
• Confirm removal of partition with typing "Yes" and press Confirm. 
• Reassign the disk(s) to the Cache Pool and
• start the Array. (Might be that you need to select the Yes checkbox on the right side)
• Format the Cache Disk/Pool by clicking the format Cache button that will be available to after the Array has started. 
• Restore with the rsync commando mentioned in Nxt97 post after the formatting.