https://www.filehorse.com/download-plex-media-server/old-versions/
I’m in exactly the same boat as seemingly many others…
I’m currently running on server version 1.21.3.4046 on a Windows 10 Pro machine, since seemingly the versions afterwards lead to my server crashing.
The problem ocurred the first time after updating to 1.21.4.4079, with the server crashing some times after the installation and after every restart/reboot as well.
I skipped a few versions and then applied the 1.22.1.4228 update, hoping that the issue had been resolved. Sadly it hadn’t, the server is still always crashing.
Reverting back to 1.21.3.4046 has the server running without any problems again.
I’ve attached the logs as well as the crash dumps from those faulty version, hoping someone might help me find the reason behind the crashes.
Crash Dumps.zip (171.0 KB) Plex Media Server Logs_2021-04-03_11-35-14.zip (6.0 MB)
Those crash dumps, do you get them from localappdata? I’m not getting anything in those 2 folders
Well 1.22.2.4282 went dead on me overnight. It stayed up for days but it seems that the problems aren’t resolved after all.
For anyone who’s been dealing with server crashes since the mess that was 4228, the issue is your database (odds are), basically all of the versions from 4228 - 4275 would have messed up your database, especially if you flip flopped server versions because Plex was seemingly “not working” during the database changes.
With the database changes, after updating my server took about 4 minutes before the web UI was available.
The only thing that resolved the issue for me was grabbing a backed up version of my database from before I updated to 4228 and allowing the migration to happen on it.
I did that. I thought it worked, it did for a while. I didn’t flipflop versions I only ever went forward in versions through the mess. Plex’s documented instructions for repairing database also seem out of date and don’t work, at least when trying to do so using Windows. If database was corrupted, Plex is supposed to actually detect and notify you of that and there have been no signs of it on my end. If it truly was the database, I would expect it to just not work… but while it’s up everything works fine.
It’s something occurring during the butler maintenance it seems since in my case it seems to always occur while nobody is using the server overnight.
I can’t explain why this is happening to you, especially if you didn’t flip flop like I did. Plex’s documentation as of the database change definitely needs an update but depending on what you’re trying to do their procedures do still work. I used their procedure for moving watched stats over and it worked - gave a bunch of errors, but it did work (for Linux, I’m on Linux).
You’re correct though, Plex is supposed to notify you of database corruption, but I don’t believe the database it truly corrupt - it’s just using a different schema so it may not know where things are (I’m guessing).
I would expect it to not work either if it’s a database problem as well. The indicator for me was watching the logs, filter for warning and error only and wait for a while to see if you start seeing the database slowly taking longer to do things. A couple milliseconds after a few hours turns into seconds which then gives the appearance and behavior that the server is crashing, I think it’s just taking so long to “respond” that devices simply timeout.
This is 100% speculation after having to fight with this myself for over a week.
If you have debug logging turned on, post your logs and I can take a look - at least until someone from Plex pops by the thread again.
I found the crash dumps in the “data directory”\Plex Media Server\Crash Reports
@kazz3r24 I created a new topic with my logs so far I haven’t had any luck in having someone from Plex look at them.
Try tagging a moderator, I can only imagine they’re busy - but giving them a “poke” might not hurt.
@ChuckPa maybe you can assist?
OUCH!
Not that hard! 
Sorry I can be more definitive but there is a file which is bad and making it 
I know it shouldn’t do that so it must be bad.
This is all I can get from both crashes. Sorry
Apr 10, 2021 02:50:20.274 [10887] DEBUG - Auth: authenticated user 1 as Binary Jay
Apr 10, 2021 02:50:20.275 [11257] DEBUG - Request: [192.168.1.221:53036 (Subnet)] GET /library/metadata/27634/children (7 live) TLS GZIP Signed-in Token (Binary Jay)
Apr 10, 2021 02:50:20.280 [10886] DEBUG - Auth: authenticated user 1 as Binary Jay
Apr 10, 2021 02:50:20.280 [11247] DEBUG - Request: [192.168.1.221:53038 (Subnet)] GET /library/metadata/32719/children (8 live) TLS GZIP Signed-in Token (Binary Jay)
Apr 10, 2021 02:50:20.283 [10887] DEBUG - Completed: [192.168.1.221:53034] 200 GET /library/metadata/32645/children (8 live) TLS GZIP 24ms 2991 bytes
Apr 10, 2021 02:50:20.301 [10886] DEBUG - Completed: [192.168.1.221:53036] 200 GET /library/metadata/27634/children (7 live) TLS GZIP 26ms 3914 bytes
Apr 10, 2021 02:50:20.310 [10887] DEBUG - Completed: [192.168.1.221:53038] 200 GET /library/metadata/32719/children (6 live) TLS GZIP 29ms 3928 bytes
Apr 10, 2021 02:50:42.515 [11254] ERROR - [Transcoder] [matroska,webm @ 0x303b463600] Invalid EBML number size tag 0x0a at pos 996964424 (0x3b6c7848)
Separate note:
I am standing up a new server so i have everything clean (only a few use it anyway).
I just finished getting PMS in ESXi with all the hardware (hades canyon) working (not bad for 2 weeks into using ESXi ?)
Also have this issue. Logs: Plex Media Server Logs_2021-04-11_12-02-03.zip (2.2 MB) Plex Media Server Logs_2021-04-11_11-33-57.zip (2.2 MB) Plex Media Server Logs_2021-04-10_22-28-01.zip (2.3 MB) Plex Media Server Logs_2021-04-10_11-37-41.zip (3.5 MB)
Hello I am having the same issue and here is my logs for my qnap server thanks Plex Media Server Logs_2021-04-11_05-24-39.zip (9.1 MB)
@ChuckPa If it is indeed a bad file (I assume you mean media file, not a Plex file) how can I tell from those logs which library item is actually causing the problem? There doesn’t seem to be any useful information in that line about what it’s actually using the transcoder on. I assume it’s probably chapter thumbnails it’s trying to create (I have video previews turned off)?
Would it relate to the item immediately above the error?
Apr 10, 2021 02:50:20.310 [10887] DEBUG - Completed: [192.168.1.221:53038] 200 GET /library/metadata/32719/children (6 live) TLS GZIP 29ms 3928 bytes
I queried the database for 32719 from that line assuming it’s metadata id 32719.
SELECT * FROM metadata_items WHERE id = 32719
This seems to point to a TV show in the database (using tvdb id found in the guid column) that I know I have watched every episode of completely without error, though. I also checked the media info for every episode in the season it seems to be referring to in the user_thumb_url column, and every one of them has chapter info in it already.
Thanks for your help.
BTW did you see the stuff in the other thread with the python errors?
I’m actually getting confused about this forum, do plex just take logs and not do anything? If you’re paying for plex pass at the very very least there should be some avenue for support.
Bye Plex, Emby time.
All:
I don’t know how I got pulled into a Windows thread but my ability to support here is minimal.
- We review each issue for each person
- We help solve the issues
- It was the weekend + some of us have personal lives.
- You won’t get a different level of support from Emby.
- Best of luck to you.
All I can see is that every time the Nvidia card is accessed, PMS fails.
Since this can’t be common to everyone, it must be something local but, for which I’m very sorry, I haven’t a clue as to how windows + nvidia drivers work.
Apr 10, 2021 22:29:52.473 [1176] DEBUG - [FFMPEG] - Loaded lib: nvEncodeAPI.dll
Apr 10, 2021 22:29:52.473 [1176] DEBUG - [FFMPEG] - Loaded sym: NvEncodeAPICreateInstance
Apr 10, 2021 22:29:52.473 [1176] DEBUG - [FFMPEG] - Loaded sym: NvEncodeAPIGetMaxSupportedVersion
Apr 10, 2021 22:29:52.598 [10888] DEBUG - [JobRunner] Job running: "C:\Program Files (x86)\Plex\Plex Media Server\CrashUploader.exe" "--directory=M:\Plex\Plex Media Server\Crash Reports\1.22.2.4282-a97b03fad" --version=1.22.2.4282-a97b03fad --platform=Windows "--platformVersion=10.0 (Build 19041)" --serverUuid=03df8d0d61440d6af8270b1d8ec71a5e237a1211 --userId=tokarevroma93@mail.ru --sentryUrl=https://sentry.io/api/1233455/minidump --sentryKey=xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx --vendor=Microsoft --model=x64 --device=PC --allowRetries=0
Apr 10, 2021 22:29:52.601 [10888] DEBUG - [JobRunner] Jobs: Starting child process with pid 8656
If everyone will forgive me, I will withdraw from this thread.
Sorry Chuck, my fault for dragging you into this!
Same issue here, since latest update. It’s crashing at complete random, having to manually launch plex again after each time
