Cache Pool Filling Up But Not Sure Why (SOLVED)

[zero_koop]

My cache pool is reporting 83% utilization out of (96GB used, 30GB free).  I can't figure out what is taking up so much space.  It's possible that Plex's appdata folder is simply using up all the space, but I've tried to find the folder(s) that use that much space and I can't find it.

The way I have my shares setup is like this:

• appdata: prefer cache
• several data shares like photos, music, etc.: Yes Cache
• other shares: No

When I browse my cache files I see the appdata folder (as expected) and the docker.img file (10GB).  I occasionally see files from the other data shares (music, photos, etc.) but those get moved off and if I invoke mover they move right away.  So everything seems to be working correctly, but why is my appdata folder reporting so much usage?  I have 4 dockers that utilize the appdata folder:

• CrashPlanPro
• dirsyncpro
• nextcloudpi
• plex

The first few don't use much space.  I've looked for the biggest folders within Plex and the best I can come up with is about 50GB of usage.  Is there something I can do to ensure my disk utilization is being reported accurately?

Edited July 23, 2021 by zero_koop

[Squid]

Shares tab.  Hit Compute All.  If it doesn't seem to ever finish, then just reload the tab and try it again.

[zero_koop]

Indeed, results show the only share using the cache currently is the appdata.  Using Windows Explorer to calculate the size of each folder adds up to 63GB.  But that is still quite a bit lower than the 96GB used being reported by Unraid.  Still trying to figure out how to account for the difference.  I believe I'm using RAID1, but I'm trying to confirm that...

 

Forgot to mention, using Unraid 6.9.2

[zero_koop]

@Squid I wonder if your advice in this thread also applies to me:

Quote

The default setting is RAID1, which means that every sector on each drive has to also be written to the other drive.  Since you've got a 1TB and a 250G the only way that's possible is that the net size is going to be 250G (btrfs does have a known problem with space reporting on mismatched drives, but the free space is always correct)

I am using btrfs with mismatched drives sizes (128GB and 240GB).  I still haven't confirmed which raid type I am using (can't figure out how to check, but I most likely used the default RAID1).  But UnRAID is reporting 30GB free and I can only find ~63GB of files on the cache drives.  Is it possible that my cache storage space is only ~93GB?

[JorgeB]

16 hours ago, zero_koop said:

can't figure out how to check,

Please post the diagnostics: Tools -> Diagnostics

[zero_koop]

Here is the Diagnostics.

tower-diagnostics-20210718-0827.zip

[JorgeB]

On 7/17/2021 at 3:25 PM, zero_koop said:

My cache pool is reporting 83% utilization out of (96GB used, 30GB free).

That is correct, something is using that space.

[zero_koop]

I cannot reconcile a few of these numbers.  What I don't know is if this is normal or not.  Here are some stats:

• Cache drive 1 size: 128GB
• Cache drive 2 size: 240GB
• filesystem: btrfs
• reported "size" on Main screen: 184GB
  • what is this about?  Shouldn't it just be 128GB?
  • as per below: 98+29.5= ~128GB
• reported "used" on Main screen: 98GB
• reported "free" on Main screen: 29.5GB
• reported utilization for drive 1 in the popup system warnings: 97%
  • how is this calculated?  98/128 = 76% or 98/184 = 53%, neither of which are close to 97%

What if I told all of my shares to NOT use the cache drives, then invoked mover?  Would that clear them out and would doing that be helpful in diagnosing the situation?

 

Just to confirm my earlier calculations, I browsed to the "cache" network drive, used Windows Explorer to calculate the size of everything in the folder and the total was 55.6GB.

Edited July 22, 2021 by zero_koop

[glennv]

If i want to see exactly what is using up space , its best to do that on unraid itself from command level.
use the following will show you a drilldown of the directory sizes at highest directory level. Go to deeper directory levels when you find what is reported as eating up the space. This will also include hidden files etc.
du -h -d1 /mnt/cache

to list files and sizes including hidden files in a directory
ls -lah /mnt/cache/etc/etc

[ChatNoir]

5 hours ago, zero_koop said:

What if I told all of my shares to NOT use the cache drives, then invoked mover?  Would that clear them out and would doing that be helpful in diagnosing the situation?

That would not work, as the Mover does not touch shares set to No or Only.

You would have to go to Settings, set the VM Manager and Docker to Disable.

Set all share to Cache : Yes

Run the Mover.

 

You can see from there what's left on Cache, for example with Midnight Commander (mc in the terminal).

[JorgeB]

6 hours ago, zero_koop said:

reported "size" on Main screen: 184GB

• what is this about?  Shouldn't it just be 128GB?

 

This is normal with different size devices, used/free will be correct.

[zero_koop]

21 hours ago, glennv said:

If i want to see exactly what is using up space , its best to do that on unraid itself from command level.
use the following will show you a drilldown of the directory sizes at highest directory level. Go to deeper directory levels when you find what is reported as eating up the space. This will also include hidden files etc.
du -h -d1 /mnt/cache

to list files and sizes including hidden files in a directory
ls -lah /mnt/cache/etc/etc

Thank you so much, I found the culprit!  And thanks to everyone else as well, I really appreciate it.  The files that were taking up space were in my Deleted Files bin in Nextcloud!  Roughly 26GB of deleted files.  Nextcloud is relatively new to me so I didn't know it was keeping my files like that.  I need to look for a setting that auto-deletes after 30 days, but otherwise I'll have to remember to empty my recycling bin. 😜