Unraid crashes, maybe because of Plex

[sarus]

Hello.

I tried to find the solution to my problem in the forum, on internet... But nothing

It seems that the Plex (official) container is crashing all my server.

I don't find anything in the logs (but I'm not an expert...).

I'm using only cache drives, I'm on the latest version of the container. I'm using hardware transcoding.syslog-192.168.0.11.logPlex Media Server Logs_2023-03-07_19-45-47.zip It's working well, when it does not crash...

I hope someone could help me to find the solution.

Thank you very much.

[horridwilting]

Hey! I also have had two odd crashes in the last week with my previously-rock-solid-for-6+-months server. 

Only things that have changed:

 

1) Updated the firmware on my cache pool 2x980 PRO SSD's (not using this Cache Pool for anything Plex, but it DOES have the appdata on it for Docker...)

2) Updated Plex to the newest version (as of like a week ago, I noticed there is an "update available" but I need to get numbers - currently at work so I'll get those as soon as I can.


Both crashes were very strange. Like, the server didn't fully reboot, but the array was stopped and it noted UNCLEAN SHUTDOWN. After the first one, I started everything back up, and did a parity check. It ran for about 20hrs then randomly crashed again...but weirdly, it sent a notice that "parity check finished" despite still having about 8hrs left on it. So, something is doing a weird..."soft" reset of some kind? 

I did see a MCE in the SYSTEM LOG but need to look at it more and probably make my own post with my DIAGNOSTICS.

For anyone to help on here, they'll ask for your DIAGNOSTICS so you should add that to your OP when you have time.

[sarus]

Hello.

Well yes I forget that ! Here they are.

For your information, last night I stopped plex container and untill now, no chrash ! It really seems to be because of plex but I still not have idea and I don't really know how to find information with all the logs.

homer-diagnostics-20230308-2007.zip

[horridwilting]

What version of Plex were/are you running? I was on 1.31.2.6739 using BINHEX-PLEXPASS and now I just forced an update to Version 1.31.2.6783 since it was available, though I can see this is a BETA version so I may need to roll that back to something earlier. I am going to keep my Plex Docker OFF for a while longer to see if it crashes or not though, just to make sure this isn't some other obscure hardware problem and the Plex thing was a red herring.

Seems the PUBLIC version is .6782
https://www.reddit.com/r/PleX/comments/11mmq4y/new_public_pms_version_available_1311678277dfff442/

[sarus]

I'm running the offical one : plexinc/pms-docker, always up to date.

[sarus]

Hello.

I've tried several things and the issue is still here. I can confirm this is plex.

I've installed netdata to try to see something before the crash but I did not see anything.

Please help 😢.

[horridwilting]

Hey sarus, sorry I didn't reply sooner - since my last post on March 9, I haven't had any more crashes after updating my Plex. I am on 1.31.2.6783 and haven't had any crashes since then.

[sarus]

Hello.

Ok thank you. i'll try binhex version and I tell you.

[sarus]

Hello.

I've tried binhex last version and... it crashed again

I don't know what to do and what to check.

When I look in NetData I just can see an increasing number of cpu interrupts. I also see that the container is increasing highly cpu utilization just before the crash and also the memory page fault.

[horridwilting]

Ah, that really sucks. Sorry.

I would check the UNRAID SYSTEM LOG (TOOLS>SYSTEM LOG). Scroll down to the bottom then go up, and see if you find any errors, especially hardware or CPU or RAM. they'll be in RED for easy ID. That hopefully will tell you what is faulting to narrow it down.
 

[sarus]

Well I can find a lot of "kernel: PMS LoudnessCmd[15154]: segfault at...". If I'm not wrong it can be because the ram is full. I've 64G of ram and it's used at 30% all the time. I suspect Plex to do something during the night and maybe fulling the ram. My problem is I can't determinate what is this f**ing thing.

[horridwilting]

Oh! You might temporarily disable SCHEDULED TASKS or the MAINTENANCE tasks in Plex just to see if that activity is causing the issue as it normally happens overnight.

[sarus]

Hello.

I've just disactivated all tasks I could but it's not possible to fully disable. I'll see if it's better... But even if it is, I would prefer to understand how to manage to don't have issue when everything is enabled. I hope we'll find...

Thank's again for your idea.

[chadcb]

I'm having this issue also. The only resolution I have at this point is to set the plex Docker template to use the following repository:
 

plexinc/pms-docker:1.30.2.6563-3d4dc0cce

 

When I first experienced the crashes I determined that it was right after I had updated Plex. So I disabled the updated Plex docker and crashes went away. And as soon as I enabled Plex, crashes started again. I then rolled back to the version listed above and ran with zero issues. So I've been stuck on that version for awhile now.

Yesterday I decided to try it again and upgrade Plex to the current version. This morning I noticed my server crashed again. So for now I just set the version back to what I was using until I can figure out the issue.

[sarus]

Hello.

Thank you for sharing. I am still having crashes even with all tasks disabled.

How could it be possible seriously !! I'm very surprised that only you and me encounter this problem... I hope someone will arrive on this topic with the brilliant idea !

 

@chadcb can you tell me if you are using hardware transcoding ? Just to know if we have similar configs.

[chadcb]

1 hour ago, sarus said:

Hello.

Thank you for sharing. I am still having crashes even with all tasks disabled.

How could it be possible seriously !! I'm very surprised that only you and me encounter this problem... I hope someone will arrive on this topic with the brilliant idea !

 

@chadcb can you tell me if you are using hardware transcoding ? Just to know if we have similar configs.

 

Yes, I'm using Nvidia. Here is what the Nvidia Driver plugin shows:
 

Nvidia Driver Version: 470.141.03

NVIDIA GeForce GTX 760

Unraid Version: 6.11.5
Kernel: 5.19.17-Unraid

Open Source Kernel Module: No

[sarus]

Hello.

Thank you @chadcb it seems ok now but how is it possible !! I hope I'll find help on plex forums.

Did you try to open a topic on this ?

[chadcb]

On 4/21/2023 at 4:00 AM, sarus said:

Hello.

Thank you @chadcb it seems ok now but how is it possible !! I hope I'll find help on plex forums.

Did you try to open a topic on this ?

 

Sorry for late reply. I haven't opened a topic yet. Also I tried updating Plex recently and the crashed started again. So reverted back.

[thellamapaul]

I also have this issue, I also use an nVidia GPU for transcoding (750 Ti). Identified Plex as the issue after changing the time scheduled tasks start and the reboot time followed. Since I've been tracking it, it happened on 4/9, 4/14, 4/28, and 5/1, so pretty sporadic. I have Plex log to my Syslog and attached the logs around the most recent 3 incidents. According to the logs it looks like thumbnails, but there are plenty of times thumbnails are generated without the crash. I'm guessing it's something else that doesn't log.

plex reboot logs.txt

[Craig Dennis]

Unraid 6.12.1

Plex 4.1.08 (official)
 

My server was rebooting every hour (or there about) and it used to be every day (for the last week) and I realised I’d changed the library scan interval in Plex.

 

I ran unraid in safe mode and turned off Docker etc. no reboots. Enabled Docker and Plex - reboots every hour. Turned off scheduled library scan and uptime was 8 hours plus (overnight) so I’m concluding that’s the culprit. 
 

I’ve just re-enabled library scan for hourly so we’ll see if it crashes in an hour. 
 

I have a ping sensor on home assistant so I can clearly see when it went down and came back.

 

Any ideas what would cause this behaviour? Next step is a fresh Plex install in a new container. 

[ich777]

On 3/9/2023 at 6:17 PM, sarus said:

Well yes I forget that ! Here they are.

Have you yet tried to switch to IPVLAN in the Docker settings or are you using IPVLAN already?

[Craig Dennis]

2 hours ago, Craig Dennis said:

I’ve just re-enabled library scan for hourly so we’ll see if it crashes in an hour. 

No crash yet after 3 hours (over 2 since enabling hourly scan) so maybe not Plex library scan. I’ll enable a new Docker every couple of hours to see if it happens again. 

[Craig Dennis]

Ok it seems like every time the server transcodes something it crashes. 
 

After several hours of stability I tried to watch something on my phone, opened Plex Dash app and it was a spinning wheel. 
 

Then I got a notification saying the server was offline (from Home Assistant). I wonder if this is a CPU issue or a RAM issue triggered by Plex transcoding. It does transcode, and shows hw in the details. This time it stayed off instead of rebooting. 
 

I’m not sure why it was crashing every hour before and disabling/enabling the scanner fixed that. 
 

More testing needed. 

Edited June 25, 2023 by Craig Dennis

[Craig Dennis]

Unraid is now crashing every night at around 3am (corresponding to the scheduled maintenance window - periodic scanning disabled). I'm going to change that and see if the reboots change as well.

Currently running:

Unraid 6.12.2
Plex 4.108.0

I have ordered a new power supply to rule that out.

Edited July 5, 2023 by Craig Dennis

[Joshndroid]

I have the same issue.

Random, what I believe, is Plex Docker Crashes.

It seems to cause the system to hang and fall off the network.

It is still powered on, but i can't get into it and it doesn't respond to a power button press (to initiate a safe shutdown)

Only a forced power cycle fixes it. This then triggers a parity check.

I have C-States disabled already in my x99 system 5820k, 32Gb ram, SSD cache drive

i have a Nvidia 1650 as my Transcode pass-through GPU.

This has been a somewhat regular occurrence for months.

 

Edit - I am already running IPVLAN

Edited July 14, 2023 by Joshndroid