For background, I have the following pools/array:

- cache_nvme (1x WD SN770 1TB)

- cache_ssd_mirror (2x TeamGroup T-Fforce Vulcan Z 1TB)

- vm_nvme (1x WD SN770 1TB)

- Main array (140TB with dual parity, 15 drives either 8TB or 14TB)

 

I had originally set up using Space Invader's recommendations - VM's use the vm_nvme, AppData on the cache_nvme, and various other files I wanted fast access to on the cache_ssd_mirror.

 

I then had the bright idea to assign AppData secondary storage to the main array, then I also started to put my media files in a folder that happened to be on the cache_ssd_mirror without realizing it.  I started getting the utilization warnings/errors on the ssd mirror while copying files over.

 

I've removed the array assignment on AppData so it only points to the nvme cache, and placed my media share on the array only.

 

But now I have a couple issues:

- I have media files clogging up the SSD cache that should be on the array

- I have AppData files on the array that should be on the nvme cache

 

I'd like to know how I can move these around so I can clean it up.

 

I have also started to get corrupted SQLite DB errors in Plex and other apps that use it.  I was reading that the mix of array/nvme on the SQLite DB files is likely a cause of the issue there, so if I get the storage assignment cleaned up, I should be able to recover the DB files affected (I've done it a couple times already, but it keeps recurring).

 

Some of the media files on the SSDs are already on the array and appear to be duplicates.  I'm guessing I can just safely delete them if so, but if there are any that are missing from the array, I need to move them but don't want to mess up where UnRAID thinks they are if I start moving stuff around on individual drives.

 

For bonus points, I have two 8TB drives that are almost 90% full, but several empty drives.  I've used Unbalanced to empty a drive in preparation for removal, but is there a way to only partially move files around to more evenly distribute the files?  Or should I just not worry about it?

 

Thanks very much for any help/advice!  Sorry for the wall of text!

 

Attach Diagnostics to your NEXT post in this thread.

Sorry, didn't include since I didn't think it was relevant.  Attached now,

nasbox-u1-diagnostics-20250316-1926.zip

8 minutes ago, SickPup404 said:

didn't include since I didn't think it was relevant

Easier to understand, less likely to misunderstand, and more complete than any description you might give.

 

 

 

The only pool that is very full is cache_ssd_mirror. The only shares that have files on that pool are isos and M---y. And those are the only shares configured to put files on that pool.

 

Just to make sure you haven't made the mistake of putting files at the top level of that pool (and so not included in any share):

What do you get from the command line with this?

ls -lah /mnt/cache_ssd_mirror

 

Good morning,

 

First, just want to thank you for your help - it's greatly appreciated!

 

This is the output from the ls command:

root@NASBox-U1:~# ls -lah /mnt/cache_ssd_mirror
total 16K
drwxrwxrwx  1 nobody users  26 Mar 17 04:40 ./
drwxr-xr-x 25 root   root  500 Mar 17 07:25 ../
drwxrwxrwx  1 nobody users  58 Mar 15 14:19 M---y/
drwxrwxrwx  1 nobody users  10 Mar 16 20:40 data/
drwxrwxrwx  1 nobody users  24 Feb 11 15:40 isos/

 

The data folder is where the media files are.  When I open data in File Manager, these are the subfolders (data is the top one):

 

The sequence I originally did on the data share:

• I had the array initially configured as primary
• Got that bright idea to make cache_ssd_mirror the primary with the array as secondary
• Got the utilization warning/errors while copying media files
• Changed the data share back to have array as primary and removed cache_ssd_mirror

As I've been looking deeper, the files that are on cache_ssd_mirror in the data share are not dupes and I need to move them off somehow.  Last  night I:

• Opened cache_ssd_mirror in File Manager
• Selected the data folder
• Tried to Copy it to a different folder NOT under the data share (/mnt/user/data) elsewhere on the array (/mnt/user/newfolder)
• Got an "Invalid Target" message.

I do have a mobile drive dock that I'm currently copying the files over to.  After that's complete, I just need to be able to delete them from cache_ssd_mirror and put them back on the array without messing something up.  Is it as simple as that - delete them from cache_ssd_mirror, then copy them from the mobile drive back to the data share on the array?

 

The other issue with AppData (files on the array when it is assigned to the cache_nvme pool) was picked up by Fix Common Problems over night.  I'll follow recommendations there to resolve that.

 

Thanks!

For mover to take any action you need both a Primary and Secondary storage location defined and the Mover Direction to what you want.

 

in the diagnostics it shows no files for the data share on the pool which is a bit strange, but with the correct share settings they should get moved.

So I should:

• Reassign the data share to have cache_ssd_mirror as prinary with the array as secondary
• Wait overnight and Mover will take care of it

I think since I removed the cache_ssd_mirror after it filled up, but before Mover could run, that caused the issue.

 

THANK YOU!

1 hour ago, itimpi said:

no files for the data share on the pool

d--a                              shareUseCache="no"      # Share exists on disk1, disk2, disk3, disk5, disk6, disk7, disk8, cache_ssd_mirror

I missed that one for the same reason, the pool is listed last. When the diagnostics were taken, it wasn't configured to put files on cache_ssd_mirror or any other pool, so I guess it was changed when it started filling up.

 

46 minutes ago, SickPup404 said:

• Reassign the data share to have cache_ssd_mirror as prinary with the array as secondary
• Wait overnight and Mover will take care of it

Yes it will probably take a few hours to move that much data.

 

When it's finished, post new diagnostics. There is still more work to be done.

 

1 hour ago, trurl said:

There is still more work to be done.

Good to know!  I will post back after Mover runs.

OK, ready for next steps.

 

Completed:

• Reassigned data share to have cache_ssd_mirror as primary, array as secondary
• Triggered Mover in the Scheduler
• Changed data share back to array only as primary
• Verified files got moved (manually picked up one straggler that didn't get moved)
• Deleted empty folders in the data folder of cache_ssd_mirror

Attached current diags.

 

Thanks!

 

nasbox-u1-diagnostics-20250317-1425.zip

Ideally, appdata, domains, and system shares would have all files on pool, with nothing on the array, so Docker/VM performance will not be impacted by slower array writes and reads, and so array disks can spin down since these files are always open.

 

Here is the current situation with these shares:

appdata is set to only put files on cache_nvme, and not move. It currently has files/folders on cache_nvme and on disks 1, 8 which will not be moved

domains is set to put files on vm_nvme, and move them to the array. It currently has files/folders on vm_nvme and disks 1, 8.

system is set to put files on cache_nvme, and move them to the array. It currently has all files/folders on disk1.

 

isos share is usually only needed to install VMs. If you put it on the array, you may need to have your VMs unmount it when not needed, to allow array disks to spin down. It is set to put files on cache_ssd_mirror and not move. It currently has file/folders on cache_ssd_mirror, and on disk1 which will not be moved.

 

You can see all this in your diagnostics in the shares folder. The shareDisks.txt file summarizes where each share's files are.

 

Nothing can move open files. You have to disable Docker and VM Manager in Settings before you can work with these shares.

 

You should set Minimum Free for each of your pools.

https://docs.unraid.net/unraid-os/manual/storage-management/#minimum-free-space-for-a-pool

 

You also need to consider Minimum Free for each user share.

https://docs.unraid.net/unraid-os/manual/shares/user-shares/#minimum-free-space

All the shares I mentioned already have Minimum Free set, perhaps to a calculated value that you might want to edit. I didn't check the other shares.

 

I see you already have appdata backup plugin installed. If you have any important VMs you might want to consider backups for those too.

 

How do you want to use each of the pools you have?

This is great info!  I will post back to you later this evening when I have a chance to sit and think about what I intend.  (Late afternoon here - eastern US.)

 

You don't realize how much I appreciate the expertise!  (I know just enough to be dangerous to my server!)

Ok,

 

Here’s my thoughts, please feel free to recommend changes where needed.

 

POOLS

vm_nvme

I’d like to use this solely for VM’s since it’s the fastest of the drives.  At this point, I only have one running, but plan on a couple others.  They are not critical, but I would like to have them backed up since it can be such a pain to reinstall/reconfigure everything.

 

cache_nvme

I’d like to use this solely for AppData, again because of speed.  As you noticed, I do have the AppData Backup plugin installed.  I’ll look over it to ensure it’s backing up as I intended.

 

cache_ssd_mirror

My thought was to have this as the main storage for things like My Documents from Windows.  I may not understand completely, but I was hoping that it could be configured so that anything done during the day just remained on the SSDs during the day, then overnight, Mover could move it to the array.  If I access something that is less frequently used, it would pull from the array, but more frequent stuff would stay on the SSDs (if that makes sense).

 

 

SHARES

AppData

As noted above, I’d like this to be on the cache_nvme pool and not on the array.  Would be backed up.

 

Domains

As noted above, I’d like this to be on the vm_nvme pool and not on the array.  Would be backed up.

 

 

ISOs

Currently doesn’t have any files in the share, and no mappings in any VMs.  Could really be put on any pool, just not sure which is ideal - maybe vm_nvme?

 

System

I’m not sure exactly what is on here.  I see libvirt and docker ima files.  I guess they are the vdisk files for the VMs & Containers?  Again, it can be anywhere but I’m not sure where.  Does it need to be backed up?

 

 

MISC

Ideally, I would love for the drives to spin down.  Right now, they obviously don’t.

Also, any thoughts on the bonus from my initial post (any need to worry about the distribution of data)?  My current drive utilization has two of the 8TB drives over 75%.  I've done some searches and it seems like it's nothing to worry about.

 

 

Hopefully, this helps you get me on the right path to getting this straight!

Good morning,

 

Had a busy few days at work, but was finally able to make changes for the better (I hope).  Not sure if I have something configured incorrectly, but it seems like the array isn't spinning down.  I just made another change with removing mapped drives in Windows so it only has cache pool drives mapped.

 

Anyway, latest diags are attached.

 

Cheers!

nasbox-u1-diagnostics-20250320-0908.zip

Your system share has all of its files on disk1. This share is the most important for allowing your disks to spindown.

 

appdata would have open files for specific running containers, domains would have open files for specific running VMs.

 

system share has open files when Docker or VM Manager is enabled in Settings. It is configured to be moved to cache_nvme, but nothing can move open files.

2 minutes ago, trurl said:

system share has open files when Docker or VM Manager is enabled in Settings

In those settings, you can see those are using a path in system share.

There ya go!  That orange triangle bit me on that one too (wanted to get rid of all of them when I set up).

 

I missed moving those files...  VM/Docker Managers shut down, files moved, restarted the managers.

 

Waiting for the timeout now, but I have no doubt I'm straight now!

New diags attached...  Drives still aren't spinning down.

 

All are set to default delay, and that's 30 minutes.

 

The "Shares" applet on the Dashboard shows three shares with stream and they are all cache only.

 

The Active Streams app has two - both on cache.

 

No Plex streams.

 

Where else can I look to find the culprit(s)?

nasbox-u1-diagnostics-20250320-1430.zip

Nevermind...  I had left a program open with files open on the array.  Closed it and drives will spin down now!

 

(Also fixed that the "system" share was set to cache_nvme and not vm_nvme.)