Server Version#: 1.19.5.3006
Player Version#: N/A
Crashing on CentOS 7.8 within 1 minute of service restart, and immediately after install. DMP file has characters which are unreadable on both linux and OSX but I can post it anyways if needed.
edit: Fixed by completely stopping PMS & starting again.
PlexDump.tar (1.0 MB)
gdb dump file.Logs directory under /var/lib/plexmediaserver (drill down). make a tar.gz or a zip and attach that file here.I’ve already resolved this but I’ve edited my original post with the dump files in case it helps to figure out what happened.
Thanks.
Since you had the dump files, it presume you don’t have Crash Reporting enabled?
Reporting them upstream automatically sends the DMP file to Plex where Engineering can look at it.
It’s very difficult for us to setup such an environment.
I do have crash reporting enabled and unfortunately it’s crashing again. Hopefully all those crash reports are getting where they should. Is there anything I can do to resolve this before a fix is released? Is there a way for me to find the previous version of PMS Beta?
May I see the ZIP of the Logs directory?
I can probably find out what’s happening.
If you look, 1.19.4 is still public.
Here’s the entire Logs directory. Thank you so much for taking a look at this, you’re legendary. The tar file appears to be a bit too big for the forum at 35 megabytes so I’ve uploaded it on DropBox. https://www.dropbox.com/s/ecdd5wnsxniyimq/PlexLogsDir.tar?dl=0 or change dl=1 to download the entire tar.
Doesn’t look like those crash dumps are being uploaded, it’s missing an API key for sentry.
Thanks. I have them. You may remove the link now if you wish.
I’ll start reading.
I can get the debug symbols and, with your minidump file, see where it’s crashing.
I do see the crash uploader being called and trying to upload but failing due to a what looks like a network error.
I also see a lot of curl timeouts (both error 28 and 60)
I am curious;
br0 showing on LAN.Which is the preferred network adapter? (it shouldn’t cause a crash but might explain the other things I’m seeing).
I can get the debug symbols and dig into this deeper but it appears 1.19.5 has been removed from PlexPass beta. I will therefore presume Engineering found something.
eno1 is the WAN adapter going to the internet and br0 is the LAN adapter going to switches at home. The public WAN IP shouldn’t be used on br0 or any LAN adapters, only on eno1. Here’s the log from messages when it crashes:
Jun 30 00:05:25 systemd: Started Plex Media Server.
Jun 30 00:06:21 sh: *** Error in `/usr/lib/plexmediaserver/Plex Media Server’: free(): corrupted unsorted chunks: 0x00007fe02c169f60 ***
Jun 30 00:06:21 kernel: Plex Media Serv[25630]: segfault at b8 ip 0000000000a6eaef sp 00007fe057ffe7b0 error 4 in Plex Media Server[954000+dda000]
Jun 30 00:06:21 systemd: plexmediaserver.service: main process exited, code=killed, status=11/SEGV
Jun 30 00:06:21 sh: ****** PLEX MEDIA SERVER CRASHED, CRASH REPORT WRITTEN: /var/lib/plexmediaserver/Library/Application Support/Plex Media Server/Crash Reports/1.19.5.3006-da9510f6b/PLEX MEDIA
SERVER/56e9f739-961f-689f-0e790b79-4880bf74.dmp
Jun 30 00:07:51 systemd: plexmediaserver.service stop-sigterm timed out. Killing.
Jun 30 00:07:51 systemd: Unit plexmediaserver.service entered failed state.
Jun 30 00:07:51 systemd: plexmediaserver.service failed.
Jun 30 00:07:57 systemd: plexmediaserver.service holdoff time over, scheduling restart.
Jun 30 00:07:57 systemd: Cannot add dependency job for unit firewalld.service, ignoring: Unit is masked.
Jun 30 00:07:57 systemd: Stopped Plex Media Server.
Jun 30 00:07:57 systemd: Starting Plex Media Server…
do you have br0 set as the preferred interface in Plex and default route in Linux?
Error 4 :
[chuck@lizum targets.508]$ perror 4
Exit code(4): error "Interrupted system call"
[chuck@lizum targets.509]$
I’ve not seen that kind of error in a long long time.
No, the default route is eno1 (unless the destination is 192.168.2/0 , then the route is br0) as per the routing table.
Kernel IP routing table
Destination Gateway Genmask Flags Metric Ref Use Iface
default lo0-100.NYCMNY- 0.0.0.0 UG 102 0 0 eno1
98...* 0.0.0.0 255.255.255.0 U 102 0 0 eno1
172.17.0.0 0.0.0.0 255.255.0.0 U 0 0 0 docker0
192.168.2.0 0.0.0.0 255.255.255.0 U 425 0 0 br0
Trying to check the Plex settings but it crashes every few seconds so just waiting for that window.
Ok… so that’s good.
Were you able to download 1.19.4 again?
In general, we always recommend keeping a stable package (public version) around to fall back to in case a beta does have issues.
1.19.4 is indeed back again on the download page! I will start keeping a copy of the previous rpm around. If anyone on the Plex team debugging needs any insight on this please feel free to share the logs you grabbed off of me (as long as it’s not public). 19.4 seems stable so far. Thanks again Chuck.
For anyone in the same position, the command is yum downgrade.
I’ll find out what happened when Engineering is back in the office tomorrow.
Same issue on 1.19.5.3021. Crashed overnight and crashes every few seconds as well. Reverted back to 1.19.4.2935 for now
Manually grab the logs before the other is gone please.
1.19.5.3006 was pulled because of Windows.
This crash is easily replicated so I’ll be uploading crash logs on Monday. I forgot to turn off my auto update so waking up this AM had the same crashes happening on 1.19.5.3035. Would you want the dmp files as well or just the logs?
This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.
