After finding Plex down I went to my dockers to find the service didn't start.

The docker log states: 

Error starting daemon: lstat /var/lib/docker/volumes/95af7c5856d1f2b33fa4ad0d01f739bdd5222425290aada3c411dc27de63a59c: input/output error

 

I have attached my diagnostics files

tower-diagnostics-20180914-1941.zip

Edited September 17, 20187 yr by ucliker

Docker image is corrupt, delete and recreate, see the docker FAQ if you need help how to.

I tried to do that but it won't let me. I keep getting Btrfs errors on my cache drive, parent transid verify failed on 1275643199488 wanted 2588463 found 2587150.

 

I copied what I could off of the drive, first I thought it was a bad cable but after switching cables and even pulling the drive and putting it in my Archlinux box it still throws the same errors. 

I'm on a phone and can't see the diagnostics now but if the cache itself also has transid errors it needs to be reformatted

Yeah, that's what I figured. Thanks for your input though. Btrfs doesn't handle power outages well and apparently, my old UPS has a bad battery. 

I want to reformat the drive to xfs but I can't find any instructions other than how to replace a cache drive with a new drive. 

 

 

To reformat a drive you need to stop the array: click on the disk to be formatted and change the format to the one you want; restart the array and the disk will now show as unmountable.   You can now use the option to format unmountable disks.

 

NOTE:  this will erase any existing data on the drive so make sure you first copy anything you want to keep to another location.

3 minutes ago, itimpi said:

To reformat a drive you need to stop the array: click on the disk to be formatted and change the format to the one you want; restart the array and the disk will now show as unmountable.   You can now use the option to format unmountable disks.

 

NOTE:  this will erase any existing data on the drive so make sure you first copy anything you want to keep to another location.

Thanks! Yeah, I copied what I could but the filesystem on the cache drive is corrupted. 

How do I format the cache drive to xfs? Under cache settings, my options are Auto, BTRFS and BTRFS Encrypted. When I manually format to xfs then add the cache back the system wants me to reformat and it reformats back to BTRFS.

How many drives do you have assigned to the cache?    If it is more than 1 then the only option will be BTRFS as XFS Is only an option if the cache is limited to 1 drive.

1 minute ago, itimpi said:

How many drives do you have assigned to the cache?    If it is more than 1 then the only option will be BTRFS as XFS Is only an option if the cache is limited to 1 drive.

I only have 1 cache drive

5 minutes ago, ucliker said:

I only have 1 cache drive

How many cache slots are showing?

Change cache slots to 1

Just now, jonathanm said:

How many cache slots are showing?

With the drives mounted only the 1 slot

Just now, johnnie.black said:

Change cache slots to 1

after stopping the array

That fixed it thanks! 

Hi everyone

 

I might be in a simulair situation

 

I wanted to change my Cache to btrfs encrypted....

I then copied all data to my UAD drive and copied it back again and now get the: "Docker Service failed to start."

 

Did I do something wrong:

1) Stop docker service

2) Copy appdata (Not system) back to the Cache drive

3) Start docker service again

 

UPDATE: Did above again 2 times and then it worked (Some transfer error?)

 

 

 

 

 

 

 

Edited January 28, 20206 yr by casperse

On 9/15/2018 at 7:05 PM, JorgeB said:

Change cache slots to 1

 

We have one german user with a degraded RAID because his SSD in slot 1 became defective. Now his docker service does not start anymore. Is it really needed to convert the RAID back to single and move the running SSD to slot 1 (as long his waiting for the new SSD to re-create an RAID1).

 

I wonder because in both cases would be the path /mnt/cache/system or /mnt/user/system or is the docker.img mounted in a different way?

44 minutes ago, mgutt said:

We have one german user with a degraded RAID because his SSD in slot 1 became defective.

If the pool was redundant you can have just slot2 populated, if it's not working there are other issues.

9 minutes ago, JorgeB said:

If the pool was redundant you can have just slot2 populated, if it's not working there are other issues.

 

It looks like Unraid mounted the cache pool read-only. This would explain his problem. We suggested him to make a backup through rsync first and then try to re-assign the SSD as a single cache drive in slot 1 if he don't want to wait for the replacement SSD.

2 minutes ago, mgutt said:

try to re-assign the SSD as a single cache drive in slot 1

Very unlikely that will help.

1 minute ago, JorgeB said:

Very unlikely that will help.

 

No? Isn't it possible to downgrade the RAID1 to Single disk through this way?

16 minutes ago, mgutt said:

No? Isn't it possible to downgrade the RAID1 to Single disk through this way?

I guess you mean re-format? If that's it then yes, but it can be left in slot, that won't matter, backup cache and re-format, I assume it this post:

Looking at the diags there were errors on the remaining device:

 

Jan 20 09:12:47 Unraid-Tower kernel: BTRFS info (device nvme0n1p1): bdev /dev/nvme0n1p1 errs: wr 1314931, rd 1043949, flush 26985, corrupt 1428, gen 0

 

This suggests that device also dropped offline in the past and it was never corrected, when the other device failed it was corrupt, tell the user to look a look here for better pool monitoring.

 

 

8 minutes ago, JorgeB said:

I guess you mean re-format?

No, I hoped there is something implemented like "btrfs balance start dconvert=single". ^^