Hi, I have read a few posts on this but my cache disk is stubbornly refusing to empty. I want to replace it because it has a lot of errors.

"Warning [Unraid Cache disk SMART health [187] SERVER] - uncorrectable error cnt is 1131"

This pops up from time to time and the mover seems to stop. I have 235 GB on the Cache disk and Mover seems to run for a few minutes, then I get these errors and although it says the mover is running, disk activity stops. If I reboot, maybe I cna move 1-2 Gb before it fails again, but this is not a tenable solution

I guess my cache disk is fubar and that is the problem.

I also cannot run my docker so can't use Krusader to move anything. Any ideas please?

Thank you all

I followed the instructions here: 

 

server-diagnostics-20250513-0948.zip

The log is showing i/o errors so I guess I cannot move anything else off my cache?

 

May 13 10:10:01 Server shfs: copy_file: /mnt/cache/system/docker/docker.img /mnt/disk2/system/docker/docker.img.partial (5) Input/output error

Is there any way to preserve my docker or at least some way to keep some of the settings that my docker apps were using if I have to start from scratch?

Losing my Plex library/history would be a bummer

Thank you

Edited May 13, 20251 yr by AngelEyes

Cache device is failing, copy what you, you can also try using ddrescue:

https://forums.unraid.net/topic/46802-faq-for-unraid-v6/page/2/#findComment-525075

 

I appreciate this option, thank you.

However if I am not worried about losing some of my media (I am not) would just replacing the cache disk and restoring the appdata backup get me where I would like to be? Or am I missing something?

Thank you

2 minutes ago, AngelEyes said:

would just replacing the cache disk and restoring the appdata backup get me where I would like to be?

You would get a new docker.img file created on the replacement disk, so you would need to reinstall the binaries for all docker containers which you can do with their settings intact using Apps->Previous Apps.

55 minutes ago, itimpi said:

You would get a new docker.img file created on the replacement disk, so you would need to reinstall the binaries for all docker containers which you can do with their settings intact using Apps->Previous Apps.

Ok, but I'd lose how everything was setup? or would restoring the appdata backup fix that?

Sorry for the basic questions

11 minutes ago, AngelEyes said:

or would restoring the appdata backup fix that?

The appdata backup restores each containers working (non-binary) files.   Using Apps->Previous Apps reapplies any settings done via the container’s settings page at the time it was installed.

 

you therefore need both to get back to where you were.

 

Ok, I retrieved what I could, and replaced the failing SSD.

In the Pool Devices section I can select the new device to replace the old one, but it asks for a Pool name. I don't know what to put here, AFAIK there was no Pool name, just the device name. Any ideas?

I tried calling it Cache but it says I cannot use the same name as a share.

Edited May 13, 20251 yr by AngelEyes

7 minutes ago, AngelEyes said:

I tried calling it Cache but it says I cannot use the same name as a share.

You have some other things that need cleaning up as well.

 

On the User Shares page, click the CLEAN UP button.

 

Then post new diagnostics.

Where is the 'User Shares' page? Do you mean the 'Shares' in the main menu?

As far as I can tell, I cannot start the array even if I set the cache disk to no device, therefore I can't access the share menu?

EDIT: Ok I started array without the cache disk (oops) and clicked on clean up 

Thanks.

server-diagnostics-20250513-1448.zip

Edited May 13, 20251 yr by AngelEyes

59 minutes ago, AngelEyes said:

In the Pool Devices section I can select the new device to replace the old one, but it asks for a Pool name. I don't know what to put here, AFAIK there was no Pool name, just the device name. Any ideas?

You cannot replace a pool device like that, click the pool, then "remove pool", then add a new pool with the same name, or a different one, and assign the new device.

1 hour ago, AngelEyes said:

I tried calling it Cache but it says I cannot use the same name as a share.

We need to investigate why you have a user share named Cache. Did you intentionally create a user share with that name? Probably you have some docker host volume creating that. It looks like it exists on many array disks.

 

Post a screenshot of the User Shares page.

Ok, here is the screenshot.

After clicking clean up the replaced cache disk worked just fine without me changing anything about Pool names. I think it just saw it as a replacement. 

I am trying to put the shares back as they were, currently moving App data, system, domains, downloads and 'cache' to the cache pool. I also have no idea why I have a share called 'cache' I made this machine a very long time ago and I don't recall a reason for that, but I am a novice so your guess is as good as mine.

Thanks for the help

Click Compute... for the Cache user share and post the results

Done, and attached.

not sure if you needed the updated diags but also attached.

server-diagnostics-20250513-1606.zip

So only 551MB in the user share named Cache. Do you know what it is?

No, it looks like a lot of tmp files. There are no folders or files with normal filenames.

Delete those files, then delete the user share named Cache. If it gets recreated you need to investigate what is specifying that path, such as a container host volume.

I can't delete the files. If I manually select a few they delete but selecting all and deleting doesn't work, nothing happens. There are over 3200 files

Go to Tools - New Permissions and run it on only the user share named Cache. Be sure you don't select the cache disk instead.

Completed but I can still only delete a few at a time manually.

How are you doing the delete?

 

In your screenshot above, you should be able to select all files by clicking the checkbox in the gray header line at the top of the file list, to the left of 'TYPE'

Yes, I select this, scroll to the bottom click delete from the bottom menu and in the pop-up click 'start'. The screen refreshes and zero files are deleted. Could it be that the Mover is still running? (nothing is happening but it says it is running)

I will reboot and try again.

Edited May 13, 20251 yr by AngelEyes

Looks like that user share is configured to be moved to cache, but according to those last diagnostics it didn't have any files on cache.

 

Post new diagnostics

Here it is

server-diagnostics-20250513-1746.zip