Comskip leading to stuck shows and massive memory leaks

I recently setup Plex DVR using a recently purchased HDHomeRun CONNECT DUO receiving free OTA broadcasts in southern Utah. I’m running the latest official plex docker container (currently 1.16.0.1226).

After enabling commercial skipping I noticed random episodes get stuck spinning under my activity dropdown in Plex. They are stuck in this state until I restart the Plex docker container which causes the show to just be deleted from the .grab folder. The worst part is that while they are stuck processing there seems to be a memory leak that gets progressively worse the more episodes you have stuck. Just check out this graph of the last 30 days of my Plex instance (should never be using this much memory):

8aa80408551fc0ce87556e6a3a06298dc2e8ee0f_2_690x243

The only reason this hasn’t been a showstopper is because my home server has 144GB of ram and I am only using a small fraction of that currently which gives me enough time to restart it every couple days (I did have to cut back on how many shows I was watching because of that massive spike).

I ended up investigating this issue and found that Comskip is the culprit (by utilizing the linux top command from inside the Plex container). The full command from the stuck process running is:

/usr/lib/plexmediaserver/Plex Commercial Skipper --ini=/config/Library/Application Support/Plex Media Server/DVR/comskip.ini --output=/media/Videos/TV Shows/.grab/db642912755a899f6cf4857b5dbcfb7f5e51d104 -t --quiet /media/Videos/TV Shows/.grab/db642912755a899f6cf4857b5dbcfb7f5e51d104/Cops (1989) - S22E08 - Dangerous Arrests.ts

Using this information I navigated to the log file file inside the .grab folder for this Comskip command. I noticed that the log file is 51MB and continually growing. The last bit of the log is just these lines repeated over and over:

Frame Rate set to 119.880 f/s
DFps[1]= 59.940 f/s
RFps[1]= 59.940 f/s
AFps[1]= 59.940 f/s
Frame Rate corrected to 59.940 f/s
Frame Rate set to 119.880 f/s
DFps[1]= 59.940 f/s
RFps[1]= 59.940 f/s
AFps[1]= 59.940 f/s
Frame Rate corrected to 59.940 f/s
Frame Rate set to 119.880 f/s
DFps[1]= 59.940 f/s
RFps[1]= 59.940 f/s
AFps[1]= 59.940 f/s
Frame Rate corrected to 59.940 f/s
Frame Rate set to 119.880 f/s
DFps[1]= 59.940 f/s
RFps[1]= 59.940 f/s
AFps[1]= 59.940 f/s
Frame Rate corrected to 59.940 f/s
Frame Rate set to 119.880 f/s
DFps[1]= 59.940 f/s
RFps[1]= 59.940 f/s
AFps[1]= 59.940 f/s
Frame Rate corrected to 59.940 f/s
Frame Rate set to 119.880 f/s
DFps[1]= 59.940 f/s
RFps[1]= 59.940 f/s
AFps[1]= 59.940 f/s
Frame Rate corrected to 59.940 f/s
Frame Rate set to 119.880 f/s
DFps[1]= 59.940 f/s
RFps[1]= 59.940 f/s
AFps[1]= 59.940 f/s
Frame Rate corrected to 59.940 f/s
Frame Rate set to 119.880 f/s
DFps[1]= 59.940 f/s
RFps[1]= 59.940 f/s
AFps[1]= 59.940 f/s
Frame Rate corrected to 59.940 f/s
Frame Rate set to 119.880 f/s
DFps[1]= 59.940 f/s
RFps[1]= 59.940 f/s
AFps[1]= 59.940 f/s
Frame Rate corrected to 59.940 f/s
Frame Rate set to 119.880 f/s
DFps[1]= 59.940 f/s
RFps[1]= 59.940 f/s
AFps[1]= 59.940 f/s
Frame Rate corrected to 59.940 f/s
Frame Rate set to 119.880 f/s
DFps[1]= 59.940 f/s
RFps[1]= 59.940 f/s
AFps[1]= 59.940 f/s
Frame Rate corrected to 59.940 f/s
Frame Rate set to 119.880 f/s
DFps[1]= 59.940 f/s
RFps[1]= 59.940 f/s
AFps[1]= 59.940 f/s
Frame Rate corrected to 59.940 f/s
Frame Rate set to 119.880 f/s
DFps[1]= 59.940 f/s
RFps[1]= 59.940 f/s
AFps[1]= 59.940 f/s
Frame Rate corrected to 59.940 f/s
Frame Rate set to 119.880 f/s
Ticks per frame = 4
DFps[4]= 59.940 f/s
RFps[4]= 59.940 f/s
AFps[4]= 59.940 f/s
Frame Rate corrected to 59.940 f/s

Retry=0 at frame=128649, time= 1797.34 seconds

(full log at: https://www.dropbox.com/s/fktj9qma75fk9t0/Cops%20(1989)%20-%20S22E08%20-%20Dangerous%20Arrests.log?dl=0)

I tried searching for any information regarding this and only found a forum thread from 2005 on the comskip forums that said it would be fixed so that wasn’t very helpful: http://www.kaashoek.com/comskip/viewtopic.php?f=5&t=1880

I’ve been trying to find someone else having the same issue on the Plex forums but I’m not having much luck (I posted at Memory leak since 1.15.x on 300+ Plex Server Instances in hopes that I wasn’t alone).

I copied everything from the .grab folder so that I can provide it if needed.

EDIT: I also found this post which matches my issue Show has recorded, but COMSKIP processing has not completed but I do not have the donator version of comskip to try this out (and I am not sure if it is the exact same issue).

You can use the non-donator version of Comskip to test - it won’t be multicore, but it will work. Be sure to put the comskip.dictionary and comskip.ini into the same folder with the renamed comskip.

I’m trying to get this done.

Do I have to compile comskip from source? I am running in a docker container so this makes it a little annoying.

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.