Cache weirdness / possibly my cluelessnes

[tiny-e]

I'm running into issues with (what I believe is) my cache pool.  

 

Not sure if this is related, or if I'm just lucky, -but the other day I installed a docker container with handbrake, and set it convert a bunch of old videos to a more modern format to make Plex Media Player happy.  Output was to a folder in my main media share.  Not too long after, one of the VMs that I run started auto-pausing itself.   Some googling points to a full cache drive/pool.  

 

My cache pool currently is comprised of 2 non-identical SSD drives I had laying around.   a 240 GB and a 512 GB drive.   Unraid GUI reports the cache pool as being 376 GB, which would be have the size of the drives.  Looking at the drives in a terminal, I see:

root@Nimbus:/mnt# du -sh cache/

212G cache/

 

root@Nimbus:/mnt# btrfs fi df /mnt/cache

Data, RAID1: total=221.56GiB, used=220.83GiB

System, single: total=4.00MiB, used=48.00KiB

Metadata, single: total=2.01GiB, used=525.73MiB

GlobalReserve, single: total=184.39MiB, used=0.00B

 

I ran the mover script (it's still running) and now my VM will start.. (note the above terminal output is after mover has been running for maybe an hour)   is it likely that all the converted video (like 200 half-hour videos) clogged up the cache drive and were waiting to be moved out?   (why didn't they move out?)

 

Would using larger/symmetrical drives for my cache pool save me from the issues that I've run into today?  

 

ALSO -- I decided to move the plex home folder from appdata to a different location on the array --would this have contributed to the issue (I didn't realize the VM was borked when I did this).

 

I've ordered 2 1TB SSDs to replace my current cache setup, any tips on an easy way to migrate from what I have to the 1TB setup?

 

TIA for any insight you can provide. 

 

 

Unraid Version: 

Version 6.7.2 2019-06-25

Base distro:

kernel-firmware: version 20190620_7ae3a09

php version: 7.2.19 (restore sqlite support)

sqlite version: 3.28.0

Linux kernel:

version: 4.19.56

 

Diagnostic info attached. 

 

 

nimbus-diagnostics-20191211-0151.zip

[JorgeB]

5 hours ago, tiny-e said:

Unraid GUI reports the cache pool as being 376 GB

This is a known btrfs issue when using different size devices, usable size will be 240GB.

 

5 hours ago, tiny-e said:

System, single: total=4.00MiB, used=48.00KiB

Metadata, single: total=2.01GiB, used=525.73MiB

Also because of this bug your cache in not redundant, there are instructions to fix it in that report.

 

 

[tiny-e]

Thanks everyone for the help.  I got plex to run.  Decided to try to move plex stuff off of the cache drive, created other weirdness.   Trying to avoid the nuke/pave option, not sure if I'm gonna make it. 

 

[tiny-e]

New cache drives arrived today.  Moved files off of cache to another drive.  Then moved them back to the new cache pool.  Docker complained about needing to rebuild the image.   So I let it.  BUT it appears to now be creating the docker.img in /mnt/user as well as /mnt/cache .

 

Docker settings show /mnt/cache/docker.img as the path for the image. 

Not exactly sure why that is. 

 

nimbus-diagnostics-20191217-0437.zip

[tiny-e]

Just now, BRiT said:

/mnt/user includes /mnt/cache, so it's not actually 2 different files.

Why would it be showing up in both places...  Unraid is hardlinking one?

[tiny-e]

Also... VMs that I deleted have just reappeared.  WTF? 

[tiny-e]

And I mean deleted days and reboots ago.. 

 

[tiny-e]

So it's normal for the docker image to be there too?   I honestly don't remember seeing it there before.

 

root@Nimbus:/mnt/user# ls -la

total 26214432

drwxrwxrwx  1 nobody users          84 Dec 17 09:21 ./

drwxr-xr-x 14 root   root                  280 Dec 17 09:11 ../

drwx------  1 nobody users                43 Dec 14 17:35 .Trash-99/

drwxrwxrwx  1 nobody users          70 Nov 14 12:02 MacBackup/

drwxrwxrwx  1 nobody users          4096 Dec 13 16:11 Media/

drwxrwxrwx  1 nobody users          8 Dec 10 13:05 PlexMedia/

drwxrwxrwx  1 nobody users          103 Dec 16 04:40 UNRAID_cache_backup/

drwxrwxrwx  1 nobody users          78 Dec 17 09:22 VMs/

drwxrwxrwx  1 nobody users          126 Dec 16 19:24 appdata/

-rw-rw-rw-  1 nobody users             26843545600 Dec 17 16:09 docker.img

lrwxrwxrwx  1 root   root             14 Nov 26 17:21 domains -> /mnt/user/VMs//

drwxrwxrwx  1 nobody users          14 Dec 16 19:22 system/

root@Nimbus:/mnt/user# 

[JorgeB]

7 hours ago, tiny-e said:

So it's normal for the docker image to be there too? 

You have it configured to be there:

DOCKER_IMAGE_FILE="/mnt/cache/docker.img"

 

New one will recreated if it doesn't exist.

[tiny-e]

10 hours ago, johnnie.black said:

You have it configured to be there:

DOCKER_IMAGE_FILE="/mnt/cache/docker.img"

 

New one will recreated if it doesn't exist.

Here's what I see...  I've re-done it a couple of times to make sure it wasn't a glitch.  

[JorgeB]

That's normal, everything in /mnt/cache/ is also in /mnt/user/, it's the same file.

[tiny-e]

OK -- Thanks.  I may just not have noticed it from before. 

Edited December 18, 2019 by tiny-e