Hi all,

 

I have been having issues with Plex on my local UnRAID box on my LAN.

 

I started looking into it today and deleted and recreated the Plex docker and it all worked for 10 mins then said Live TV could not be played. I drilled into the logs and get an error message repeatedly showing;

 

PMS: failure detected. Read/write access is required for path: /config/Library/Application Support/Plex Media Server

 

So I looked at the array and worryingling several shares are showing as Cache = Prefer: Cache or Yes: Cache.

 

I say worryingly because the 'cache' drive is a very old 1Tb Apple disk which I was only using to store CCTV video on from one of my VMs and never intended to be used as a cache disk in the truest sense and this morning I got a cache disk warning saying it was "read only" - this didnt worry me as it was supposed to only be CCTV footage! But now I can see several shares (appdata / domains / system / isos / downloads / VMs) showing Prefer:Cache or Yes:Cache

 

I have never really used cache disks, so read on the forum that in order to move share data from the cache drive to change Cache setting to "Yes: Cache", invoke mover (to move back on to the protected array disks?) and once moved to set cache to No.

 

So, I tested this on a not so important share (VMs - dont ask )  - once I set cache to yes and invoked mover I can look at the share and where it originally showed all data on "cache" it now shows some on cache and some on disk 3;

 

 

Whenever I go back to mover it shows MOVE NOW - Click to invoke mover... and if I click MOVE NOW it says its moving, but as soon as the page is refreshed it shows the MOVE NOW button again.

 

Am I being too impatient - or am I dowing something wrong?

 

Any help to get my data back into a protected environment before the "cache" disk fails would be very welcome.

 

Thanks

You need to disable docker/VM services first, then run the mover, if it still doesn't move everything after that enable the mover logging and post the diags after running the mover one more time.

Thanks - will do it now and report back

3 hours ago, JorgeB said:

You need to disable docker/VM services first, then run the mover, if it still doesn't move everything after that enable the mover logging and post the diags after running the mover one more time.

 

Thanks @JorgeB

 

I have done this and only after re-running it a few times some shares are now showing as just on protected drives and so I have set cache on them to No.

 

But some shares still show data on the cache drive - after it finishes, I invoke mover again, it says its running but nothing seems to change... e.g.;

 

 

If I keep re-running it will it eventually get everything off the cache disk?

8 minutes ago, SliMat said:

If I keep re-running it will it eventually get everything off the cache disk?

Unlikely,

3 hours ago, JorgeB said:

if it still doesn't move everything after that enable the mover logging and post the diags after running the mover one more time.

 

OK, it ran again but still not everything has cleared, so am running again with logging enabled.

@JorgeB Attached is the diags after invoking mover again...

unraid-diagnostics-20221213-2042.zip

Filesystem went read-only, reboot, run the mover, post new diags.

2 minutes ago, JorgeB said:

Filesystem went read-only, reboot, run the mover, post new diags.

 

Thanks - I saw that too and wondered whether it was reporting that this failing disk is now read only?

 

I will do this now and post new diags in a bit.

 

Thanks

New diags attached @JorgeB

 

There are a number of VM images and backups which I can delete as they are no longer needed - but at the moment some of them show as on cache+diskX - so I wasnt sure whether to delete these to create more space, or if as they are partially moved it would create an issue?

 

Thanks

unraid-diagnostics-20221214-1608.zip

Cache device appears to be in a very bad state, it's still moving data but likely won't be able to save everything.

Thanks @JorgeB

 

The data exposed to the cache disk seems to be dockers appdata and some VM's so looks like I will have to delete all these in order to be able to remove the cache disk and then start again.

 

I may see what docker appdata & VMs can be moved onto the array and then just delete everything and start again

 

I guess that the /mnt/user/appdata/system folder is not related to dockers and is needed for the array??

31 minutes ago, SliMat said:

I guess that the /mnt/user/appdata/system

Should be /mnt/user/system/appdata, that's used for docker containers.

Mine is set to /mnt/user/appdata (also set to this on my production machine too);

 

that's fine, you mentioned /appdata/system, and that would be unusual, but possible.

Ah, I thought I saw a folder /mnt/user/appdata/system - but you're right its not there

 

So, to clear this fault it looks like my appdata, domains and VMs shares have lost data;

 

 

So would the best way forward be to remove the cache drive and then delete these folders, create an empty /mnt/user/appdata folder and recreate my dockers? Obviously when I remove the cache drive the system will (probably) show these 3 shares as all OK because there will no data on a cache disk as this will no longer be there.

 

5 minutes ago, SliMat said:

create an empty /mnt/user/appdata

No need since that folder will already exist, change it to cache=prefer and run the mover after the new cache is installed, if you want to see if some of the moved data can be saved, if not just delete the existing folder.

I'm just removing the cache disk for now as I never set it up intentionally, it was a really old SATA Apple disk which was only supposed to store CCTV video - so I wasnt worried that it was a very old dubious disk and have no idea when/how it became included in the system at all

 

Or, do you mean I need to add a new cache disk, set cache=prefer, run mover, then remove the cache disk?

 

In the long run I plan to add a SSD as a cache disk to host VMs on and then I'll take advice on what other data is worth having on there (dokers, etc?) and how to set it up properly - including a backup

You don't need to have a cache device. 

Thanks - will try it today 🤞