Hello,

I have an issue where my cache is filling because the mover isn't moving as I believe I have it set to. I do not nor have I ever installed the mover tuning plugin. The cache has filled and is not being used per my settings to not exceed 90% capacity, but it will not move on my nightly schedule. 

 

I have attached how my shares are configured, and a copy of the logs.

 

Any thoughts or ideas would be appreciated. Thank you

Edited April 14, 20251 yr by zerickk

Attach Diagnostics ZIP to your NEXT post in this thread. 

tower-diagnostics-20250414-1640.zipThank you. here is the diagnostics

https://docs.unraid.net/unraid-os/manual/shares/user-shares/#default-shares

 

Ideally, these shares would have all files on cache or other pool, with nothing on the array, so Docker/VM performance won't be impacted by slower array reads and especially writes, and so array disks can spin down since these files are always open.

 

appdata is the only one of these shares with any files on cache, but it also has files on the array. And it isn't configured to move anything currently, which is just as well for now since we need to make room on cache for all of these shares.

 

d----r                            shareUseCache="only"    # Share exists on cache
p--x                              shareUseCache="no"      # Share exists on cache, disk1, disk2, disk3

These are the only other shares with files on cache, and they aren't currently configured to be moved to the array.

 

What is the purpose of these shares?

 

I assume you don't really want p--x to keep any files on cache, since you have it configured to not write to cache currently. But it isn't configured to move any of those cached files to the array either.

 

What do you get from command line with this?

du -h -d 1 /mnt/cache

 

Thank you.

 

Valid point for the shares setup. I don't actually have any VM's or docker data in domains so it should be ok as cache --> array. But i have set appdata as cache only and system as array only. I believe that should be ok?

 

Docker is my docker image and is set to cache only. I have plex set for cache --> array to mitigate daily writes and speed write time every time new movies are added. I want plex to load to the array with mover nightly.

 

du -h -d 1 /mnt/cache command provides this"
60G     /mnt/cache/docker

 

I have attached a photo of my current cache disk files. It is the plex data that is filling it with content that isn't moving to the array. 

 

Thanks again for your assistance. 

25 minutes ago, zerickk said:

I don't actually have any VM's or docker data in domains so it should be ok as cache --> array.

Then you don't really need the domains user share at all.

 

25 minutes ago, zerickk said:

But i have set appdata as cache only and system as array only. I believe that should be ok?

No, for the reasons I have already given.

45 minutes ago, trurl said:

Ideally, these shares would have all files on cache or other pool, with nothing on the array, so Docker/VM performance won't be impacted by slower array reads and especially writes, and so array disks can spin down since these files are always open.

appdata as cache-only would be OK if all of it were actually on cache, but it has files on the array. And system should be all on cache as well.

 

25 minutes ago, zerickk said:

Docker is my docker image and is set to cache only.

I assume you mean docker and not Docker. Linux is case-sensitive, so these are not the same.

 

But, according to your docker.cfg, your docker.img is at

DOCKER_IMAGE_FILE="/mnt/user/system/docker/docker.img"

so apparently not where you think it is.

 

It is in system share, which is the standard location for it. But as noted, system share is all on the array instead of all on cache where it should be.

25 minutes ago, zerickk said:

du -h -d 1 /mnt/cache command provides this"
60G     /mnt/cache/docker

That doesn't agree with the screenshot, which shows appdata and plex also on cache. Maybe you didn't let the command complete. Try the command again.

 

25 minutes ago, zerickk said:

I have plex set for cache --> array to mitigate daily writes and speed write time every time new movies are added. I want plex to load to the array with mover nightly.

But you don't have plex configured to be moved at all. And it isn't configured to write anything to cache.

 

You should install Fix Common Problems plugin.

 

Also, you have 25 .cfg files in config/shares that don't correspond to any user shares. Probably you accidentally created a lot of user shares, maybe by writing to the top level of disks or pools (any folder at the top level is automatically a user share). Then you removed all those shares, but their .cfg files are still hanging around. You can clean that up (and clean those out of your diagnostics) by clicking the CLEAN UP button on the User Shares page.

 

Let's start by getting plex and that unused (useless) docker user share moved to the array.

 

Set plex and docker user shares to Primary:cache; Secondary:array; Mover action:cache->array.

 

Run mover, wait for it to complete. This may be most of the 400G on cache, so it will probably take many hours.

 

When it finally completes, post new diagnostics.

 

I will check back tomorrow.

6 minutes ago, trurl said:

Then you don't really need the domains user share at all.

Looking at the link you provided, I should move my VM from a different SSD to this location and set it as cache only to properly use this share. I will fix that.

 

8 minutes ago, trurl said:

No, for the reasons I have already given.

I think I understand this now. I should have both on either cache or array not a mix like I do now? I can set both to cache when I can sort out the space issue.

 

10 minutes ago, trurl said:

I assume you mean docker and not Docker. Linux is case-sensitive, so these are not the same.

Thank you, yes I mean docker with lowercase

 

11 minutes ago, trurl said:

so apparently not where you think it is.

Thank you, I didn't realize this either. You are pointing out many configuration issues I have here and I appreciate it. I see that the docker.img i have in the "docker" cache only share hasn't been modified for some time and is clearly not current. To rectify this, I would delete that "docker" share and move the "system" share to cache to align with the issue above where my "appdata" is on cache?

 

14 minutes ago, trurl said:

That doesn't agree with the screenshot, which shows appdata and plex also on cache. Maybe you didn't let the command complete. Try the command again.

You are right. Here is the full result:

60G     /mnt/cache/docker
127G    /mnt/cache/appdata
209G    /mnt/cache/plex
396G    /mnt/cache

 

15 minutes ago, trurl said:

But you don't have plex configured to be moved at all. And it isn't configured to write anything to cache.

Is this not configured to write to cache and move to array?

 

16 minutes ago, trurl said:

You should install Fix Common Problem plugin.

Will do. Thank you.

 

17 minutes ago, trurl said:

Also, you have 25 .cfg files in config/shares that don't correspond to any user shares. Probably you accidentally created a lot of user shares, maybe by writing to the top level of disks or pools (any folder at the top level is automatically a user share). Then you removed all those shares, but their .cfg files are still hanging around. You can clean that up (and clean those out of your diagnostics) by clicking the CLEAN UP button on the User Shares page.

Done. Thank you.

 

11 minutes ago, trurl said:

Set plex and docker user shares to Primary:cache; Secondary:array; Mover action:cache->array.

They already are set this way. In the past this would move with mover and keep my cache from filling, but now mover isn't moving them.

 

Thanks again! I have used unraid for 6 years now and see that I have a fair bit of configuration improvement opportunities.

25 minutes ago, zerickk said:

I think I understand this now. I should have both on either cache or array not a mix like I do now?

No, you should have appdata and system shares all on cache with nothing on the array, 

1 hour ago, trurl said:

Ideally, these shares would have all files on cache or other pool, with nothing on the array, so Docker/VM performance won't be impacted by slower array reads and especially writes, and so array disks can spin down since these files are always open.

 

25 minutes ago, zerickk said:

They already are set this way.

Looks like there was probably a "duplicate" plex.cfg, probably actually named Plex.cfg, and the user share was using the wrong one.

 

Since you have cleaned that up, post new diagnostics and I will see what more needs to be done about that share configuration.

1 minute ago, trurl said:

No, you should have appdata and system shares all on cache with nothing on the array, 

Ok both on cache

 

I added the new diagnostics file. 

 

Thank you!

 

tower-diagnostics-20250414-2054.zip

32 minutes ago, zerickk said:

Ok both on cache

The screenshot shows how you have them configured. To find out where the files actually are you can click Compute... for each share and wait for the result.

 

If you had installed Fix Common Problems, it would be telling you the shares location doesn't match its settings.

 

Making appdata and system shares cache-only means they won't be moved at all, since they have no Mover action.

 

But we are getting ahead of ourselves, we don't want them moved yet until we make room on cache for them. So just leave them like that for now.

 

Your new diagnostics does show you have cleaned up all those extra .cfg files in config/shares as I suggested, except...

P--x                              shareUseCache="yes"     # Share does not exist
p--x                              shareUseCache="no"      # Share exists on cache, disk1, disk2, disk3

As suspected, you have a Plex.cfg. The Plex user share is configured to be moved to the array, but it doesn't actually exist since there is no top level Plex folder.

 

And the plex user share, which does exist (because it is a top level folder), doesn't have a corresponding .cfg file, so it hasn't actually been configured and won't be moved.

 

You need to get rid of that Plex.cfg, then you can make settings for plex.

 

Dynamix File Manager is built-in to V7, you can browse to the config/shares folder on flash by going to MAIN - Boot Device, and clicking on the ↗️ arrowed square icon to the left of Flash to browse into /boot, then clicking on config to browse into config, then clicking on shares to browse into shares. Then you can select Plex.cfg and Delete it.

 

Then go to the User Shares page and click on plex to get to its settings, and set it to

Primary:cache; Secondary:array; Mover action:cache->array.

 

Then post new diagnostics so I can confirm we are ready to move things from cache.

 

10 hours ago, trurl said:

The screenshot shows how you have them configured. To find out where the files actually are you can click Compute... for each share and wait for the result.

 

If you had installed Fix Common Problems, it would be telling you the shares location doesn't match its settings.

 

Making appdata and system shares cache-only means they won't be moved at all, since they have no Mover action.

 

But we are getting ahead of ourselves, we don't want them moved yet until we make room on cache for them. So just leave them like that for now.

Ok I installed the Fix Common Problems plugin and see that there are some items needing attnetion.

 

10 hours ago, trurl said:

As suspected, you have a Plex.cfg. The Plex user share is configured to be moved to the array, but it doesn't actually exist since there is no top level Plex folder.

 

And the plex user share, which does exist (because it is a top level folder), doesn't have a corresponding .cfg file, so it hasn't actually been configured and won't be moved.

 

You need to get rid of that Plex.cfg, then you can make settings for plex.

 

Dynamix File Manager is built-in to V7, you can browse to the config/shares folder on flash by going to MAIN - Boot Device, and clicking on the ↗️ arrowed square icon to the left of Flash to browse into /boot, then clicking on config to browse into config, then clicking on shares to browse into shares. Then you can select Plex.cfg and Delete it.

 

Then go to the User Shares page and click on plex to get to its settings, and set it to

Primary:cache; Secondary:array; Mover action:cache->array.

Ok the Plex.cfg issue makes sense to me now. I went in and deleted it and then reconfigured the plex user share to Primary:cache; Secondary:array; Mover action:cache->array.

 

Here are the new diagnostics.

 

Thank you again!

tower-diagnostics-20250415-0741.zip

appdata                           shareUseCache="only"    # Share exists on cache, disk2
domains                           shareUseCache="yes"     # Share exists on cache, disk1, disk2, disk3
d-------s                         shareUseCache="yes"     # Share exists on cache, disk1, disk2, disk3
n-------d                         shareUseCache="yes"     # Share exists on cache, disk1, disk2, disk3
p--x                              shareUseCache="yes"     # Share exists on cache, disk1, disk2, disk3
system                            shareUseCache="only"    # Share exists on disk1, disk2

These are the shares we are concerned with. Mover will move all the "yes" shares to the array. We don't want domains to be a "yes", so set it to

Primary:cache; Secondary:none for now until we make room on cache for it and others.

 

Looks like we have some other shares on cache that weren't there before, so lets do this again.

 

What do you get from command line with this?

du -h -d 1 /mnt/cache

 

 

33 minutes ago, trurl said:

These are the shares we are concerned with. Mover will move all the "yes" shares to the array. We don't want domains to be a "yes", so set it to

Primary:cache; Secondary:none for now until we make room on cache for it and others.

Here is my new setup:

 

33 minutes ago, trurl said:

What do you get from command line with this?

127G    /mnt/cache/appdata
228G    /mnt/cache/plex
9.4G    /mnt/cache/downloads
0       /mnt/cache/domains
2.3M    /mnt/cache/nextcloud
0       /mnt/cache/data
364G    /mnt/cache

 

Thank you

Looks like plex + downloads + nextcloud = about 240G to be moved to the array.

 

Run mover, wait for it to complete, then post new diagnostics.

Excellent. The mover finished successfully. Here are the diagnostics. 

 

I imagine this will be the next issue I need to address?

tower-diagnostics-20250415-1135.zip

cache now has nearly 300G available.

 

downloads and nextcloud still have files on cache. Maybe you are still adding to those shares, or maybe they have some open files. We'll probably take care of open files when we disable Docker for this next part.

 

Nothing can move open files, so before you can work with those "default" shares

 

Go to Settings - Docker and disable. Go to Settings - VM Manager and disable.

 

Set appdata, domains, and system shares to

Primary:cache; Secondary:array; Mover action:array->cache

 

Run mover, wait for it to complete. appdata will probably take a while, plex appdata has lots of files.

 

Then post new diagnostics.

 

tower-diagnostics-20250416-0912.zipThank you. I have finished the mover and checked the shares. It appears everything moved properly. Also the fix common issues plugin errors cleared up. Here are the diagnostics. 

Looks good!

 

Thank you so much for all the help!