Plex Using 100+ GB of RAM during Schedule Tasks

Server Version#: 1.25.0.5282
Player Version#: 1.38.0.2734-ec4489fe
OS (of server): Gentoo
Plex Media Server Logs_2021-12-02_04-39-36.zip (4.8 MB)

For the past week, my Plex server has been crashing randomly during playback. It crashed on my 6 times on Sunday alone, and at least 2 to 3 times daily since. I’ve updated to the most recent release and it’s still crashing frequently. I’ve not changed anything in my settings and have been running this server without this issue for around 5 years to this point. Needless to say, I am confused and frustrated and would appreciate any help and insight as to what is happening and how I might be able to fix it.

Thank you.

1 Like

I’ve reverted back to version Version 1.24.4.5081 and the crashing seems to have stopped. So I guess I’m hanging in this version until Plex can get stable again with their “stable” releases.

Same here, server crashed a couple of times already, which was never happening before…

So for me, it looks like there’s a Plex memory leak somewhere. My crashes have come from my being hosted on a Whatbox instance and their fair use limiter shutting Plex down because it keeps suddenly jumping up to 100GB of RAM usage. I’m not sure why this is suddenly happening with the latest version, but it’s definitely happening. If any one knows why Plex is suddenly having moments of trying to use so much RAM I’d love to know.

TO ALL FOLLOWING HERE:

Please do the following -

  1. Settings - Server - General - Show Advanced
  • Verify DEBUG logging is enabled
  • Verify VERBOSE logging is disabled
  • UNCHECK (Disable) Send Crash Reports to Plex
  • Click SAVE
  1. Now we must wait for the next crash

  2. When it does crash

  • Make a ZIP or tar.gz of the “Crash Reports” directory (Plex Media Server/Crash Reports)
  • Start PMS & Download the Logs ZIP file
  1. Attach both the Crash Reports ZIP/tar.gz and the logs.

I will try to find out where it’s crashing.

A friend was reporting me the infamous “E4” error when casting since the last update. It seems the server crashes were occurring after this error. I’ve now reverted to the previous version and have not had a single crash since (neither the E4 error)

UGH…

While I understand your desire to keep it working.

Without logs and a crash report… I have no way of helping find out so the bug can be fixed.

Crash Reports.zip (278.1 KB)
Plex Media Server Logs_2021-12-06_04-12-51.zip (6.6 MB)

It crashed again tonight. Here are the files you have requested

Crash Reports.zip (278.1 KB)
Plex Media Server Logs_2021-12-06_14-49-26.zip (6.4 MB)

It happened again while I was sleeping. So here’s the files for that crash.

Crash Reports.zip (278.1 KB)
Plex Media Server Logs_2021-12-07_04-49-37.zip (4.6 MB)

Here’s from tonight’s first crash. I’m figuring the more data the better, right? My apologies if I’m providing too much.

Crash Reports.zip (278.1 KB)
Plex Media Server Logs_2021-12-07_05-57-53.zip (4.5 MB)

Here’s from tonight’s second crash.

Thanks for these.

I looked at them and there are no DMP files.

Did you remember to disable (uncheck) “Send Crash Reports to Plex” and it SAVE ?

FWIW, you can delete all the old empty directories

Yes, I had disabled “Send Crash Reports to Plex” and saved the settings. The problem I’m facing is that (as I had post earlier in thread) it turns out that my cloud server host (Whatbox) has a limiter script that is killing my Plex instance every time it has a sudden burst of RAM usage (jumping up to 100 GB of RAM). So I’m no longer certain that my Plex is crashing in the traditional way. But there clearly is something that’s causing a memory leak of some sort that I’ve been looking for some guidance on. My server host was suggesting that I try a fresh install with a fresh database setup (the idea there being that something in my database had become incompatible with the current versions of Plex since the database switch over), but I’m hoping to avoid that as I’ve put a lot of working into a chuck of the metadata of my library (either custom entries or corrections/edits/rematching) over the years of my using Plex. So if there’s still any insight you can help provide my situation, I would grateful. I do appreciate you taking the time to help out with this issue.

@ElusiveZatchmo

If you get into the directory structure and down to the databases, how big are they?

I’ve never seen PMS use over 100GB in the RSS value (resident segment size) – that would bust everything there. It would beat up my 256 GB server pretty hard if it were hitting 100 GB.

Do you know what the actual triggers are (from their FAQ / help docs) ?

@ChuckPa
Getting into the directories.

My database directory is 28 GB. 
My com.plexapp.plugins.library.db file is 2 GB.
My com.plexapp.plugins.library.blobs.db file is 4 GB
My com.plexapp.plugins.library.db-wal file is 2 GB
My com.plexapp.plugins.library.db-shm file is 2.34 MB

I don’t know what the actual triggers are tbh, I was getting this information from them when talking about this with their support team:

We are now seeing more information which indicates that for some reason your Plex instance is using 100 GB of RAM and is being automatically killed by the fair use system. We aren’t sure why however and continue to investigate why this continues to occur

…you obviously aren’t meaning intentionally to use 100GB but the automated system called the OOM killer regards it as a threat to the whole shared server and is instantly killing it.

… there is definitely starting to become a concern that these issues are being caused directly by incompatibilities between the way the newer versions of Plex interact with databases that were created on older versions of Plex initially.

This has been the assessment so far with what they’re able to look at. Hence their suggestion to start over again to see if that clears things up, but like I had mentioned previously, I’m hoping to avoid that if you folks might be able to provide some insight.

I’ve tried “clean bundles” and “optimizing database” in hopes that that might fix some things, but that doesn’t seem to have corrected the problem as my Plex is still being killed off because of a memory leak issue.

@ElusiveZatchmo

I’m looking at another user’s DB . His is 3.5 GB.

I think you have the same problem.

If this is true, It’s not a memory “leak” in the traditional sense.
It’s the database returning so much data for the query that the trigger fires.

If I open a PM to you , would it be possible to get a copy of your DB (pms stopped) ?

@ChuckPa yeah I can do that; just let me know which file specifically you’re looking for and if I need to leave my server off until I hear back from you. I can shoot that your way when you PM me.

@ElusiveZatchmo

PM sent

@ChuckPa
Got it and replied. Thanks

Throwing in my +1 since I’m also experiencing daily crashes on every version of 1.25. I reverted back to Version 1.24.5.5173 and things seem fine again. Good luck in finding the cause so this gets fixed ASAP.