Unmountable Cache Drives

[FattyDave18]

I have over the last month setup my UnRaid server and moved my data to it/setup a few VMs and dockers.

 

I am running 2 cache disks in what should be a Raid 1. Due to a previous issue I wanted to confirm the Raid 1 was working so while the server was powered down I unplugged the SATA cable on 1 of the cache drives and powered it on.

 

It booted up correctly and started the array with the cache drive (although I noted it reported twice the capacity for some reason).

 

I turned it off, plugged the 2nd cache drive back in and powered it up. The array started again but only with 1 drive in the cache disk pool.

I also found I wasn't able to stop the array successfully. I eventually got in to the disk settings menu disabled auto start and restarted. Once restarted I was able to add the ssd back to the cache drive list but it is now reporting as unmountable.

 

Is this to be expected? I'm not in a position where I seemed to have lost access to my data on the cache drive.

I have followed this guide: https://forums.unraid.net/topic/46802-faq-for-unraid-v6/?do=findComment&comment=543490

I am able to mount one of the drives and I can see all my files in Midnight Commander, when I try and copy them to another disk however I get the following error on all files:

Input/Output error (5)

 

Can anyone help in getting my data back?

 

Many thank in advance.

 

 

[trurl]

Go to Tools-Diagnostics and attach the complete Diagnostics ZIP file to your NEXT post 

[FattyDave18]

furnnet-servur-diagnostics-20200614-2210.zipComplete diagnostics as requested.

[FattyDave18]

I added back the second cache disk in to the cache pool and started the array. My shares now show up and I can browse to the files but still unable to copy them to another drive and docker/VMs services failing to start.

 

The drive I added to the cache pool is now listed twice, once in the cache and ones as a Historical Device.

 

Thought I'd reboot the server just in case that resolved the issue. Following restart both cache drives are now listed in the cache devices pool but as unmountable: no file system.

 

Revised diagnostics attached.

 

furnnet-servur-diagnostics-20200615-0818.zip

Edited June 15, 2020 by FattyDave18

[JorgeB]

Cache is complaining of a missing device, and at the time this happened it wasn't fully redundant, so while you should be able to mount it read only most likely some data will be missing, still you can try some recovery options here.

[FattyDave18]

Thanks for the reply and looking at my issue.

 

I get a wrong fs error when trying to mount read only.

 

I'm able to mount as degraded. I can then see all my files but when I try and copy them using mc I get the below

 

Cannot read source file "/x/........"

Input/output error (5)

 

I don't really understand how I've lost data just by unplugging a redundant drive and plugging it in again.

[JorgeB]

3 hours ago, FattyDave18 said:

I don't really understand how I've lost data just by unplugging a redundant drive and plugging it in again.

The most likely is the balance operation not finishing, either when you removed it or when it was added back.

 

19 hours ago, FattyDave18 said:

It booted up correctly and started the array with the cache drive (although I noted it reported twice the capacity for some reason).

This is normal during the balance, balance can take several minutes depending on how much data there is.

 

3 hours ago, FattyDave18 said:

Input/output error (5)

This usually means data corruption, you can use btrfs restore, but it will restore any files it can, including corrupt ones.

[FattyDave18]

47 minutes ago, johnnie.black said:

 

you can use btrfs restore, but it will restore any files it can, including corrupt ones.

 

When I run this (-v) i get:

 

~# btrfs restore -v /dev/sdc1 /mnt/disk1/restore
warning, device 1 is missing
Restoring /mnt/disk1/restore/appdata
Error mkdiring /mnt/disk1/restore/appdata: 2

 

Running with -vi I get:

 

~# btrfs restore -vi /dev/sdc1 /mnt/disk1/restore
warning, device 1 is missing
Restoring /mnt/disk1/restore/appdata
Error mkdiring /mnt/disk1/restore/appdata: 2
Restoring /mnt/disk1/restore/domains
Error mkdiring /mnt/disk1/restore/domains: 2
Restoring /mnt/disk1/restore/system
Error mkdiring /mnt/disk1/restore/system: 2
Restoring /mnt/disk1/restore/Cache Drive
Error mkdiring /mnt/disk1/restore/Cache Drive: 2
Restoring /mnt/disk1/restore/BlueIris
Error mkdiring /mnt/disk1/restore/BlueIris: 2
Restoring /mnt/disk1/restore/Downloads
Error mkdiring /mnt/disk1/restore/Downloads: 2
Done searching 

 

Nothing is copied to the restore share.

 

[JorgeB]

13 minutes ago, FattyDave18 said:

Error mkdiring /mnt/disk1/restore/appdata: 2

This suggest it's failing to create the folder.

[FattyDave18]

40 minutes ago, johnnie.black said:

This suggest it's filing to create the folder.

Ah, case sensitive... 🤦‍♂️

 

It's gone through all directories and some files do seem to have been restored OK. Others not - namely the 2 VM disk images I was most keen to saving.

 

Is it likely check --repair will fair any better or are they gone for good?

When you say it's dangerous to use I assume you mean to the data rather than to the drive or anything else?

 

Thanks again

[JorgeB]

1 hour ago, FattyDave18 said:

Is it likely check --repair will fair any better or are they gone for good?

You can try after the restore, but wouldn't keep my hopes up.

 

1 hour ago, FattyDave18 said:

When you say it's dangerous to use I assume you mean to the data rather than to the drive or anything else?

Correct.

[Kudjo]

I'm helping someone with this too (a dirty shutdown made his two SSDs in his cache pool become unmountable).

 

Did you ever figure out what to do about the "Error mkdiring" message? (JorgeB is correct, it is not creating the folders, but we're running the command as root with no luck.)

 

Any ideas?

[JorgeB]

10 hours ago, Kudjo said:

Did you ever figure out what to do about the "Error mkdiring" message?

That suggests it's restoring to an invalid path, e.g. created a folder called /mnt/disk1/Restore and are trying to restore to /mnt/disk1/restore