I am still getting the same error on my docker image after updating.
For those that want a quick python script that can reboot your docker I wrote one that reads the docker logs and restarts if it sees the error. This could easily be modified to work with a local install as well.
#!/usr/bin/env python3
import docker, time
# Set docker client and get container list
client = docker.from_env()
con_list = client.containers.list()
# Find Plex container
for con in con_list:
# Change to your container name
if con.name == "plex":
plex = con
# Stream the container log looking for the crash. Restart plex then restart the loop
while True:
time.sleep(60)
for line in plex.logs(stream=True, tail=5):
#print(line.decode("utf-8"))
# Change to the error you need to reboot on
if "double free or corruption (!prev)" in line.decode("utf-8"):
plex.restart()
May 29 19:23:35 plex sh[31581]: double free or corruption (!prev)
May 29 19:23:35 plex sh[31581]: ****** PLEX MEDIA SERVER CRASHED, CRASH REPORT WRITTEN: /var/lib/plexmediaserver/Library/Application Support/Plex Media Server/Crash Reports/1.15.8.1163-005053bb5/PLEX MEDIA SERVER/20233930-bfc4-ba55-0cacfb99-3d1c3790.dmp
My setup is using HW transcoding on a NVIDIA Quadro P2000. I also have a BTRFS SSD cache drive that contains the transcoding directory. Also my docker image is on this cache drive. The library files are stored on an XFS HDD.
This seems to happen most when I am using my 2015 Samsung that is running the old Samsung Smart TV app. It will transcode correctly, then when switching the episode or going to the next movie it fails to load, I check my docker logs and it shows the following.
Dolby, Dolby Digital, Dolby Digital Plus, Dolby TrueHD and the double D symbol are trademarks of Dolby Laboratories.
double free or corruption (!prev)
****** PLEX MEDIA SERVER CRASHED, CRASH REPORT WRITTEN: /config/Library/Application Support/Plex Media Server/Crash Reports/1.15.8.1163-005053bb5/PLEX MEDIA SERVER/04c22cd6-7cfc-71a7-7dcada58-2903c68f.dmp
Found this after Googling the “double free or corruption” error I’ve had for a few months now on a direct install on Fedora 29. I have a cronjob running every 5 mins checking systemctl status plexmediaserver for “PLEX MEDIA SERVER CRASHED” and then restarting if found. So far there has not been a pattern I can see as it can go a couple weeks without it occurring and can occur when no users are watching. Looking forward to the fix.
I have an plex installed on an unraid server and this is happening almost every day for me, no real pattern that I’ve noticed, really looking forward to a fix for this.
I expect to have a development build later today which I will be making available to try. Initially I will send it to users that were involved in trying out 1.15.5.934
You do not need to enable verbose logging. Plex specifically states to never turn that on unless specifically asked to. Just debug logging being enabled in enough.
The fix in the development is in the path that was crashing as seen in a number of dumps. Some dumps have memory corruptions and cannot be linked to this code but we think it is the same issue.
We probably need to wait a week to see if the users sent the development build have had any crashes similar to before
I have not been able to access our crash reporting system. If you get me zip of all the logs after a restart following a crash - i can check for similarities
Desperately waiting for a fix. Running an old version of Plex that still crashes due to the database maintenance issue, but at least that’s only every few days instead of every few hours.
Surely the Plex Devs can spin up an instance of Plex on their own Linux servers ton experience the issue rather than relying on customer logs.
@sa2000 can I please have the Dev build too? I’ve been experiencing this crash so often that I also have a cron script installed to restart Plex when scanning the logs for double free or corruption
edit: also sharing my script in case it’s useful to anyone, it force stops plex if the service won’t stop and it waits for the CrashUploader to finish. https://pastebin.com/NyxpunQG
which database maintenance issue ? Please point me to your forum post(s) where you reported the problem and provided logs with debug logging enabled - obtained soon after the crash
You really think the issues that users are reporting can be seen by Plex devs / support by just running Plex Media Server. It was almost impossible to repro the issue on this topic and was only seen once on a dev machine and not repeated itself again.
