Plex crashes every few weeks after scheduled maintenance period

[live_from_the_moon]

Server Version#: Version 1.19.5.3112
DSM Version#: Version 6.2.3-25426 Update 2
Model#: Synology DS918+

I’ve experienced several crashes now over the last few months. They always occur around 3-4 AM (when Plex scheduled maintenance runs), but looking over the agent logs for this last one it looks like it happened after the agent scans were done.

This last crash happened at 3:26AM, the logs only show this:

Aug 04, 2020 03:25:48.959 [0x7fb447fff700] DEBUG - Request: [192.168.0.1:5426 (Allowed Network (Subnet))] OPTIONS / (9 live) Signed-in
Aug 04, 2020 03:25:48.960 [0x7fb455ba4700] DEBUG - Completed: [192.168.0.1:5426] 200 OPTIONS / (9 live) 1ms 6611 bytes
Aug 04, 2020 03:25:52.953 [0x7fb4558b6700] DEBUG - EventSource: Failure in IdleTimeout (0 - Success).
Aug 04, 2020 03:25:52.953 [0x7fb4558b6700] ERROR - EventSource: Retrying in 15 seconds.
Aug 04, 2020 03:26:07.953 [0x7fb455ba4700] DEBUG - EventSource: Resolving plex.werehou.se port 443
Aug 04, 2020 03:26:07.954 [0x7fb455ba4700] DEBUG - EventSource: Resolved plex.werehou.se to 63.211.239.61
Aug 04, 2020 03:26:27.953 [0x7fb4558b6700] DEBUG - EventSource: Failure in IdleTimeout (0 - Success).
Aug 04, 2020 03:26:27.953 [0x7fb4558b6700] ERROR - EventSource: Retrying in 30 seconds.
Aug 04, 2020 03:26:28.132 [0x7fb4558b6700] WARN - Crash: Crash reporting disabled

Plex Media Server Logs_2020-08-04_09-37-23.zip (5.4 MB)
Appreciate any help.

[live_from_the_moon]

@trumpy81

Damaged/poorly encoded files, such as the above, are far more likely to cause a crash than anything else and during maintenance is the most likely time for that to happen.

sorry, but this makes little sense. how does a single corrupt video file crash the entire plex instance? and not even during playback, but just routine scan? that seems like a serious problem with the scanner more than anything.

[live_from_the_moon]

@trumpy81

Who said it was a single instance?

…uh, the logs do, unless I’m missing something. There’s no need to get so defensive. And arguing over semantics - scanner vs analyzer or whatever, is irrelevant, because if the process is so sensitive to errors that a single corrupt file can take down the whole server, the problem is with that process.

If Plex maintenance tools are really that fragile, though, there should be more visible notification to users rather than hoping they dig through the logs to find warnings about corrupt media, since it will apparently crash their server without warning.

[live_from_the_moon]

First, thank you for using a more balanced tone.

@trumpy81

It just seems likely since this has always been an issue and when investigated it has usually turned out to be the files at fault.

I understand, and I certainly appreciate going after low-hanging fruit from a troubleshooting perspective. And I’m not expecting Plex to moderate my media- that’s obviously the server operator’s job. But I do think that if there is a corrupt file discovered during maintenance, there should be a higher visibility.

Case in point, the file mentioned above has been there for years, and the maintenance runs every single night, and yet I’ve never even known about the issue until a few days ago, and only due to digging through multiple log files. If it really is a common cause of crashing, I think it’s reasonable and maybe even prudent for Plex to give some visual notification to the server operator about it, besides just relying on the log entries. Doesn’t have to be anything fancy, maybe just a flag in the GUI with a link to a support article, but something is needed, otherwise you can’t realistically expect people to ever know Plex has found a potentially destabilizing problem.