Docker image disk utilization of 70%

[bwnautilus]

Just noticed this warning last night.  My image file is 20GB but according to "docker system df -v" I'm only using 2.6GB.

The plex docker is using the array on which one disk just reached the 70% threshold.  Is this related to the docker warning?

 

Thanks in advance.

 

root@MediaTower:~# ls -l /mnt/user/system/docker/docker.img
-rw-rw-rw- 1 nobody users 21474836480 Aug  1 13:11 /mnt/user/system/docker/docker.img
root@MediaTower:~# docker system df -v
Images space usage:

REPOSITORY                    TAG       IMAGE ID       CREATED        SIZE      SHARED SIZE   UNIQUE SIZE   CONTAINERS
linuxserver/duckdns           latest    027cca6024b6   5 days ago     22.31MB   0B            22.31MB       1
linuxserver/nextcloud         latest    27992a0e7f58   8 days ago     420.9MB   0B            420.9MB       1
linuxserver/mariadb           latest    4ef1197eee5c   9 days ago     288.8MB   0B            288.8MB       1
plexinc/pms-docker            latest    bd30c8482314   5 weeks ago    656.4MB   0B            656.4MB       1
jlesage/nginx-proxy-manager   latest    8b2f4cf2c43f   2 months ago   191.8MB   0B            191.8MB       1
spaceinvaderone/macinabox     latest    141be9f2bd41   6 months ago   1.062GB   0B            1.062GB       1

Containers space usage:

CONTAINER ID   IMAGE                         COMMAND                  LOCAL VOLUMES   SIZE      CREATED        STATUS                  NAMES
7f899bddc97a   linuxserver/duckdns           "/init"                  0               18.1kB    4 days ago     Exited (0) 3 days ago   duckdns
250cb2242d30   linuxserver/nextcloud         "/init"                  0               50.8kB    5 days ago     Exited (0) 3 days ago   nextcloud
e512e1a314f7   linuxserver/mariadb           "/init"                  0               22.7kB    5 days ago     Exited (0) 3 days ago   mariadb
b5aa71488419   plexinc/pms-docker            "/init"                  0               11.4GB    9 days ago     Up 4 hours (healthy)    Plex-Media-Server
14497fb6d571   jlesage/nginx-proxy-manager   "/init"                  0               16.3kB    8 weeks ago    Exited (0) 3 days ago   NginxProxyManager
bea321562dfe   spaceinvaderone/macinabox     "/bin/sh -c 'bash /M…"   1               0B        6 months ago   Created                 macinabox

Local Volumes space usage:

VOLUME NAME                                                        LINKS     SIZE
21e2e215cd3e2a40baa14ec50f2d0b82e8caf434e213ef6856c41a916b62cd08   0         0B
6badc8e4b59899efac563fd45f4a3a27555d69eea5cfb3794d6908de8da41ae7   0         0B
95dba7fdb028efc2858cc810e87b8c2cdbd1bda91e92dd1f04325fba49f8dced   1         0B
cbf4d165d4969f4db849d2855c4bd049f4d21f88157ea9a3b34c459d9ecff99a   0         0B

Build cache usage: 0B

CACHE ID   CACHE TYPE   SIZE      CREATED   LAST USED   USAGE     SHARED
root@MediaTower:~# 

 

[trurl]

Attach diagnostics to your NEXT post in this thread 

[bwnautilus]

Diagnostics attached.

mediatower-diagnostics-20220801-1349.zip

[bwnautilus]

Received the warning yesterday at 12:36PM if that helps narrow down the search.

 

[bwnautilus]

Someone on the reddit unraid group was having the same problem.  Looks like the docker warning subsystem cannot distinguish between disk usage in the docker.img file vs. disk usage on the array.  Oh well.

 

Warning - Docker image disk utilization high

[trurl]

You have misunderstood that reddit. You are not using a docker folder, you have a docker.img file that is 70% full.

 

Not necessarily a problem as long as usage isn't growing.

 

Go to Docker page, click Container Size button at bottom, post a screenshot

[bwnautilus]

Container Size gives different values than "docker system df -v".  Looks like Plex is the culprit.  I'll have to go digging around in that container.

 

Thanks @trurl

 

 

Edited August 1, 2022 by bwnautilus

[trurl]

36 minutes ago, bwnautilus said:

Plex is the culprit

Where do you have Plex transcoding? Do you use DVR feature of Plex and if so where is it saving?

[bwnautilus]

1 minute ago, trurl said:

Where do you have Plex transcoding? Do you use DVR feature of Plex and if so where is it saving?

Plex transcoding is set to /tmp.  Not using DVR feature.

 

[trurl]

Just now, bwnautilus said:

Plex transcoding is set to /tmp

The screenshot shows your mapping for the  /transcode folder, but what does the plex application actually have set for transcoder directory?

[trurl]

For comparison I have 20G docker.img here is my Container Size as you can see plex uses much less than you have.

[bwnautilus]

2 hours ago, trurl said:

The screenshot shows your mapping for the  /transcode folder, but what does the plex application actually have set for transcoder directory?

Transcoder is set to /transcode which is mapped to /tmp on unraid.  From my Plex client:

 

Edited August 2, 2022 by bwnautilus

[bwnautilus]

Well this is one for the books.  Plex just released a new container build.  Installed it.  Now container size reports 660MB.  I'm suspecting previous version was foobar.

 

Thanks for your help @trurl

[trurl]

Probably simply reinstalling any version is the reason the size decreased, keep an eye on it.

 

Do you use the Download feature for watching media offline? That is another way plex might write something.

 

The usual cause of filling docker.img is an application writing to a path that isn't mapped.

 

 

[vvolfpack]

I have the same problem as OP. However, it didn't dissipate after updating the docker. I'm using the docker by Plex Inc. and have thumbnails video previews turned on. It's likely that the thumbnails are getting written directly on the image file? Uploading screenshots of my mappings within the docker and from Plex directly! Any advise would be phenomenal. Thanks!

vvolfbox-diagnostics-20221110-0807.zip

[ijuarez]

I have the same issue, using LSIO docker, I do use the DVR function and when I record 1 show my utilization ballons to 95% and typically thats the end of the show. I wish I could change it something for temp recording.

[trurl]

4 hours ago, ijuarez said:

using LSIO docker, I do use the DVR function

I have no problem with this. Where are you telling it to save recordings?

[ijuarez]

I'm presume you mean plex?  if so I did not change the default location.

Let me google that.

 

 

[Hoopster]

42 minutes ago, ijuarez said:

if so I did not change the default location.

The default Plex transcoding location is {path to Plex folder in appdata}/Library/Application Support/Plex Media Server/Cache/Transcode/

 

Even mapping /transcode to /tmp can result in your system running out of memory.  When recording with the DVR, Plex likes to use a lot of space so you can timeline scrub a show while recording.  It deletes nothing until the recording is finished and can use up to 16GB for an hour of recording.  Not ideal.  I limit the space allocated to transcoding in RAM to 16GB total (out of 64GB on my server) which forces Plex to delete session files that have already been played or recorded since I never watch a show that is currently recording.

Edited November 10, 2022 by Hoopster

[ijuarez]

14 minutes ago, Hoopster said:

The default Plex transcoding location is {path to Plex folder in appdata}//Library/Application Support/Plex Media Server/Cache/Transcode/

 

Even mapping /transcode to /tmp can result in your system running out of memory.  When recording with the DVR, Plex likes to use a lot of space so you can timeline scrub a show while recording.  It deletes nothing until the recording is finished and can use up to 16GB for an hour of recording.  Not ideal.  I limit the space allocated to transcoding in RAM to 16GB total (out of 64GB on my server) which forces Plex to delete session files that have already been played or recorded since I never watch a show that is currently recording.

Thank you for that explanation

 

 

I did find in my plex docker setup did not have a place to transcode. so.....that probably is my fix.

 

I will look at using ram to save my dvr recordings

 

[Hoopster]

3 minutes ago, ijuarez said:

I will look at using ram to save my dvr recordings

 

I have this in my go file to limit transcoding to 16GB and to recreate the folder in RAM on reboot:

mkdir /tmp/PlexRamScratch
chmod -R 777 /tmp/PlexRamScratch
mount -t tmpfs -o size=16g tmpfs /tmp/PlexRamScratch

 

Many are using as little as 4GB RAM for transcoding and it works fine.

 

With this, the mapping in the Plex docker container is /transcode to /tmp/PlexRamScratch and /transcode is specified as the transcoding location in Plex.

[vvolfpack]

On 11/10/2022 at 1:32 PM, Hoopster said:

I have this in my go file to limit transcoding to 16GB and to recreate the folder in RAM on reboot:

mkdir /tmp/PlexRamScratch
chmod -R 777 /tmp/PlexRamScratch
mount -t tmpfs -o size=16g tmpfs /tmp/PlexRamScratch

 

this is really useful! Noob question: Is there a guide that goes over how to set this up? Go File in config? 

 

[trurl]

On 11/11/2022 at 6:56 PM, vvolfpack said:

Go File in config?

yes

[ijuarez]

On 11/10/2022 at 3:26 PM, ijuarez said:

Thank you for that explanation

 

 

I did find in my plex docker setup did not have a place to transcode. so.....that probably is my fix.

 

I will look at using ram to save my dvr recordings

 

Just wanted to update

 

Since i discovered that I did not have a transacode path setup once created  I have not gotten any notifications that my docker is running out of space.
so user error.

[Snipe3000]

I'm running into this same issue. For some reason I have two transcode settings in my plex docker config. Does that mean something is wrong with my linuxserver version of plex?

 

Is there a way to clear the /tmp folder? I'm guessing this is something I have to do periodically, sense plex isnt doing it. /tmp in the root of Unraid doesnt have anything plex related in it.

Edited September 10, 2023 by Snipe3000