Non-responsive after 1.15.0.659 ubuntu upgrade

Plex Media Server Desktops & Laptops
62

if Plex Media Server fails to allocate memory - regardless of what platform it is running on - then it will log the allocation failures in the Plex Media Server log.

After the crash you can locate the log files and zip and attach with details of the problem

Please ensure debug logging is enabled beforehand - see
https://support.plex.tv/articles/201643703-reporting-issues-with-plex-media-server/

and for location of Logs - see
https://support.plex.tv/articles/200250417-plex-media-server-log-files/
and
https://support.plex.tv/articles/202915258-where-is-the-plex-media-server-data-directory-located/

63

I’ve updated PMS up through 1.15.2.793 and it still leaks memory such that I have to reboot daily. Any word on fixing this? Thanks!

64

The only known issue has already been addressed and it relates to the DVR scheduler. So the problem needs to be identified and investigated.
Suggest you run with debug logging enabled on the server. Restart the server and then at regular intervals take snapshots of how much memory is being used the Plex Media Server process. If there is a noticable spike, collect logs zip soon after noticing this and save together with the memory use snapshot and what time each was. Then when memory usage becomes excessive leading to failures, collect the logs and snapshot of how much memory is used.

This can then be investigated and to look for any clues as to where the memory may be going.

65

Sounds like a plan. Where do I post the memory snapshots and logs? Thanks!

66

Right here. Drag & drop; Allow to upload (you will see the forum respond with a URL link); his “Reply”.

67

Memory Snapshots (Complete).txt (67.9 KB)
Memory Snapshots (Jumps Only).txt (995 Bytes)
Plex Media Server Logs_2019-03-22_12-20-44.zip (3.2 MB)
PMS Logs for Memory Jumps.zip (1.3 MB)

68

The previous post has all the information I gathered, as follows:
Memory Snapshots (Complete).txt: PMS memory consumption snapshots every 60 seconds
Memory Snapshots (Jumps Only).txt: PMS memory consumption snapshots where consumption jumps.
PMS Logs for Memory Jumps: Plex rotates logs aggressively when debugging is on, so I missed some of the logs, but I gathered as many as I could.
Plex Media Server Logs_2019_03-22_12-20-44: complete logs from just before the system became unresponsive and had to be rebooted.

Let me know if you need anything more, and thanks in advance.

69

Thank you for the comprehensive set of diagnostics.

70

I have gone through your diagnostics and minute by minute memory usage and picked up 4 periods of time where we have logs and where memory jumped by about 500Mb-600Mb each time.

I have passed the evidence to the development team. It appears to be when we do the channel matching and metadata processing

This shows the times I picked

Date/Time	 	  Memory 
03/21/19 1:23	 2,754,336 
03/21/19 1:24	 3,352,356 	*****  +600Mb

03/21/19 7:05	 3,804,960 
03/21/19 7:06	 4,419,372 	*****	+615Mb	(logs available)

03/21/19 19:05	 4,394,784 
03/21/19 19:06	 4,943,660 	*****	+550Mb	(logs available)

03/21/19 23:05	 4,796,196 
03/21/19 23:06	 5,353,236 	*****	+557Mb

03/22/19 6:44	 5,508,848 
03/22/19 6:45	 5,967,600 	*****	+459Mb	(logs available)

03/22/19 10:44	 5,975,796 
03/22/19 10:45	 6,434,548 	*****	+459Mb	(logs available)
73

That’s my conclusion as well. I hope they are able to find what’s leaking. Thanks!

74

With regards to the memory snapshots - i presume these were all at :00 seconds precisely? Is that right?

Date/Time	 	 	 Memory 
03/21/19 01:23:00	 2,754,336 
03/21/19 01:24:00	 3,352,356 	*****  +600Mb

03/21/19 07:05:00	 3,804,960 
03/21/19 07:06:00	 4,419,372 	*****	+615Mb	(logs available)

03/21/19 19:05:00	 4,394,784 
03/21/19 19:06:00	 4,943,660 	*****	+550Mb	(logs available)

03/21/19 23:05:00	 4,796,196 
03/21/19 23:06:00	 5,353,236 	*****	+557Mb

03/22/19 06:44:00	 5,508,848 
03/22/19 06:45:00	 5,967,600 	*****	+459Mb	(logs available)

03/22/19 10:44:00	 5,975,796 
03/22/19 10:45:00	 6,434,548 	*****	+459Mb	(logs available)
75

My apologies, the seconds seem to have been truncated. Here they are:

Date/Time Memory
03/21/19 01:23:29 2,754,336
03/21/19 01:24:29 3,352,356 ***** +600Mb

03/21/19 07:05:32 3,804,960
03/21/19 07:06:32 4,419,372 ***** +615Mb (logs available)

03/21/19 19:05:38 4,394,784
03/21/19 19:06:40 4,943,660 ***** +550Mb (logs available)

03/21/19 23:05:51 4,796,196
03/21/19 23:06:52 5,353,236 ***** +557Mb

03/22/19 06:44:10 5,508,848
03/22/19 06:45:11 5,967,600 ***** +459Mb (logs available)

03/22/19 10:44:24 5,975,796
03/22/19 10:45:25 6,434,548 ***** +459Mb (logs available)

Let me know if you want the full corrected snapshot tables.

76

Yes please - if the data is there - could i have it with the seconds included please for both your text files

77

Here you go.

Memory Snapshots (Complete).txt (78.5 KB)
Memory Snapshots (Jumps Only).txt (1.1 KB)

78

How did you take your snapshots? I have similar problem but my usage is excessively higher. I’ve hit 4GB easily after a week of running it.

79

See this post Server Crashes While Loading EPG Data - #12 by millerim

Run a script like it after restarting the server and finding the PID fir the plex media server process

Download server logs periodically - say twice or three times a day

Once you reach high memory usage levels and the memory stays high, then capture logs and attach the memory snapshots text file ( please make sure date and time shows time including seconds ) and all the saved zipped server logs and the latest zipped logs

80

There is a variation to the script here I have got a an amendment to the script to make sure it writes to a file
See Server Crashes While Loading EPG Data to get the snapshots to a file

81

Is there an update on this memory leak? Thanks!

82

I know that the problem has been reproduced by the development team. I do not have any info yet for the fix

83

Looks like version 1.16.2.1297 still leaks, even with the new EPG system. Any word on the fix?