Plex still going down multiple times in a short period of time with 6.12.3

to add to my OP

restarted it twice while remote at a friends house and noticed in the 10min commute back home it was down again.

thats when i grabbed diagnostics and the plex logs via docker. all my libraries are unavailable and only the free plex services show like watch list and discovery

stopped array and running smart test on my 1 month old SSD however this is a unraid/zfs issue not my hardware....

prior to some btrfs issues in was having before i replaced my SSD's i hardly ever needed to restart plex now its 3-6 times per day to get it up and running...im thinking of going back to btrfs

could it be my docker.img is a btrfs vdisk on ZFS? should i wipe it and create a zfs one and re-install all my containers?

Tried docker directory and could not get my reverse proxy working at all so tried XFS vdisk and that too did not work for my reverse proxy had to go back to the btrfs vdisk image

1 hour ago, Can0n said:

restarted it twice while remote at a friends house and noticed in the 10min commute back home it was down again.

 

Maybe a ZFS-problem... my cache is running on xfs and all is fine with Plex and 6.12.3

Edited July 16, 2023 by Zonediver

15 minutes ago, Zonediver said:

 

Maybe a ZFS-problem... my cache is running on xfs and all is fine with Plex and 6.12.3

yeah i think you are right....ZFS has been a diaster for me

might have to move it all back to array and format as xfs have not use that for cache yet

Im using the offical plex contain which one are you using, i was previously using the binhex:plexpass for years with no issues other than delayed updates

17 minutes ago, Can0n said:

yeah i think you are right....ZFS has been a diaster for me

might have to move it all back to array and format as xfs have not use that for cache yet

Im using the offical plex contain which one are you using, i was previously using the binhex:plexpass for years with no issues other than delayed updates

 

I am using the plex docker from linuxserver.to - the official plex docker was not working for me...

Edited July 16, 2023 by Zonediver

Your plex hardware encoder fails.

 

Jul 15 01:04:54 Thor kernel: Plex Transcoder[976]: segfault at 148572f55000 ip 0000148576547204 sp 00007ffc1563c198 error 4 in ld-musl-x86_64.so.1[148576502000+53000] likely on CPU 12 (core 4, socket 0)

 

Test Plex without hardware encoding enabled.

 

See also https://docs.unraid.net/unraid-os/release-notes/6.12.0 about known issues with i915 driver

Changed Priority to Minor

6 hours ago, Can0n said:

to add to my OP

Please uninstall GVT-g since it seems like you aren't using it and see if this makes a difference.

 

Can you also post your docker run command from Plex?

I have a i5-10600 here and everything is running smoothly with Emby and Jellyfin everything on ZFS but without snapshots.

 

I see a lot of this messages in your syslog, are you sure that this is correct:

Jul 15 16:07:04 Thor pulseway: Periodically collected data not ready yet for Types::Reporting::Data::CoreReportData
Jul 15 16:07:04 Thor pulseway: No data for module 1
Jul 15 16:07:04 Thor pulseway: Periodically collected data not ready yet for Types::Reporting::Data::NetworksUsageData
Jul 15 16:07:04 Thor pulseway: No data for module 3
Jul 15 16:07:04 Thor pulseway: Failed to get package list
Jul 15 16:07:04 Thor pulseway: No data for module 6
Jul 15 16:12:20 Thor pulseway: Periodically collected data not ready yet for Types::Reporting::Data::CoreReportData
Jul 15 16:12:20 Thor pulseway: No data for module 1
Jul 15 16:12:20 Thor pulseway: Periodically collected data not ready yet for Types::Reporting::Data::NetworksUsageData
Jul 15 16:12:20 Thor pulseway: No data for module 3
Jul 15 16:12:20 Thor pulseway: Failed to get package list
Jul 15 16:12:20 Thor pulseway: No data for module 6
Jul 15 16:17:37 Thor pulseway: Periodically collected data not ready yet for Types::Reporting::Data::CoreReportData
Jul 15 16:17:37 Thor pulseway: No data for module 1
Jul 15 16:17:37 Thor pulseway: Periodically collected data not ready yet for Types::Reporting::Data::NetworksUsageData
Jul 15 16:17:37 Thor pulseway: No data for module 3
Jul 15 16:17:37 Thor pulseway: Failed to get package list
Jul 15 16:17:37 Thor pulseway: No data for module 6
Jul 15 16:22:53 Thor pulseway: Periodically collected data not ready yet for Types::Reporting::Data::CoreReportData
Jul 15 16:22:53 Thor pulseway: No data for module 1
Jul 15 16:22:53 Thor pulseway: Periodically collected data not ready yet for Types::Reporting::Data::NetworksUsageData
Jul 15 16:22:53 Thor pulseway: No data for module 3

 

 

I also see these lines in your go file:
 

#Hardware video Transcode Enable and TMP File System
modprobe i915
***line removed***
#chmod -R 777 /dev/dri
mkdir /tmp/PlexRamScratch
chmod -R 777 /tmp/PlexRamScratch
mount -t tmpfs -o size=32g tmpfs /tmp/PlexRamScratch

 

Can you try to not use a tmpfs for Plex and transcode somewhere on a hard disk, maybe a spinning one?

You also don't have to do the modprobe because unRAID is doing that for you and even if it wouldn't then Intel-GPU-TOP would do that for you.

 

May I also ask why are you doing that:

mount -o remount,size=1024m /var/log/

 

 

It seems that you have a lot customized maybe start pulling back a bit at least in terms of Plex and see if it does if you run it bare bones with the default transcoding directory and so on.

 

 

EDIT: Please also consider upgrading the BIOS, the BIOS that you are running (Version: 0811) is the first one that was released for this Motherboard, the newest version is: 1801

4 hours ago, ich777 said:

Please uninstall GVT-g since it seems like you aren't using it and see if this makes a difference.

 

Can you also post your docker run command from Plex?

I have a i5-10600 here and everything is running smoothly with Emby and Jellyfin everything on ZFS but without snapshots.

 

I see a lot of this messages in your syslog, are you sure that this is correct:

Jul 15 16:07:04 Thor pulseway: Periodically collected data not ready yet for Types::Reporting::Data::CoreReportData
Jul 15 16:07:04 Thor pulseway: No data for module 1
Jul 15 16:07:04 Thor pulseway: Periodically collected data not ready yet for Types::Reporting::Data::NetworksUsageData
Jul 15 16:07:04 Thor pulseway: No data for module 3
Jul 15 16:07:04 Thor pulseway: Failed to get package list
Jul 15 16:07:04 Thor pulseway: No data for module 6
Jul 15 16:12:20 Thor pulseway: Periodically collected data not ready yet for Types::Reporting::Data::CoreReportData
Jul 15 16:12:20 Thor pulseway: No data for module 1
Jul 15 16:12:20 Thor pulseway: Periodically collected data not ready yet for Types::Reporting::Data::NetworksUsageData
Jul 15 16:12:20 Thor pulseway: No data for module 3
Jul 15 16:12:20 Thor pulseway: Failed to get package list
Jul 15 16:12:20 Thor pulseway: No data for module 6
Jul 15 16:17:37 Thor pulseway: Periodically collected data not ready yet for Types::Reporting::Data::CoreReportData
Jul 15 16:17:37 Thor pulseway: No data for module 1
Jul 15 16:17:37 Thor pulseway: Periodically collected data not ready yet for Types::Reporting::Data::NetworksUsageData
Jul 15 16:17:37 Thor pulseway: No data for module 3
Jul 15 16:17:37 Thor pulseway: Failed to get package list
Jul 15 16:17:37 Thor pulseway: No data for module 6
Jul 15 16:22:53 Thor pulseway: Periodically collected data not ready yet for Types::Reporting::Data::CoreReportData
Jul 15 16:22:53 Thor pulseway: No data for module 1
Jul 15 16:22:53 Thor pulseway: Periodically collected data not ready yet for Types::Reporting::Data::NetworksUsageData
Jul 15 16:22:53 Thor pulseway: No data for module 3

 

 

I also see these lines in your go file:
 

#Hardware video Transcode Enable and TMP File System
modprobe i915
***line removed***
#chmod -R 777 /dev/dri
mkdir /tmp/PlexRamScratch
chmod -R 777 /tmp/PlexRamScratch
mount -t tmpfs -o size=32g tmpfs /tmp/PlexRamScratch

 

Can you try to not use a tmpfs for Plex and transcode somewhere on a hard disk, maybe a spinning one?

You also don't have to do the modprobe because unRAID is doing that for you and even if it wouldn't then Intel-GPU-TOP would do that for you.

 

May I also ask why are you doing that:

mount -o remount,size=1024m /var/log/

 

 

It seems that you have a lot customized maybe start pulling back a bit at least in terms of Plex and see if it does if you run it bare bones with the default transcoding directory and so on.

 

 

EDIT: Please also consider upgrading the BIOS, the BIOS that you are running (Version: 0811) is the first one that was released for this Motherboard, the newest version is: 1801

Hi thanks there is a lot its an older server for sure (over 5 years)

I swear i uninstalled GVT-g alreadt...interesting....i removed it again

I have commented out the mod probe and tmp ram drive for Plex in the go file that was how i got hardware transcoding through iQSV when I first got this server up and running back on unRAID 6.4 (i believe it was) it has always been fine until the move to ZFS.

the pulseway logs have baffled me but have been that way for 2 years just have not had the time to troubleshoot 

I suspect that were what prompted me to increase the log size to 1GB


I tried to do a force update to get a docker run displayed but it failed and created an orphan image

now i get get it installed back from apps
 

IMAGE ID [plexpass]: Pulling from plexinc/pms-docker.
IMAGE ID [a70d879fa598]: Already exists.
IMAGE ID [c4394a92d1f8]: Already exists.
IMAGE ID [10e6159c56c0]: Already exists.
IMAGE ID [d1042fe57e96]: Already exists.
IMAGE ID [ac5317c7b384]: Already exists.
IMAGE ID [47414e89d67b]: Pulling fs layer. Downloading 100% of 152 B. Verifying Checksum. Download complete. Extracting.

TOTAL DATA PULLED: 152 B

Error: failed to register layer: stat /var/lib/docker/btrfs/subvolumes/e88f3b8d60610253638260b5aecc05dc96223fc700c87f6fe077e6c1e9345215: no such file or directory



Ill take a look at the bios it is a rather newer motherboard and it was on my list of things to do

interesting enough looks like my docker.img file corrupted


all my containers went to orphaned images

 

I am deleting the docker.img file and re-doing it

Updated BIOS

Commented out the modprobe and Ram Drive

 

change log size to 512MB

deleted docker image and reset up again using docker.img in BTRFS format

re-installed all containers and set plex to use my download-cache (two spinner 4TB disks in ZFS)

Ill see how stable it is and report back


on a server reboot it typically takes less than 30 hours before i started seeing issues

On 7/16/2023 at 1:12 AM, bonienl said:

Your plex hardware encoder fails.

 

Jul 15 01:04:54 Thor kernel: Plex Transcoder[976]: segfault at 148572f55000 ip 0000148576547204 sp 00007ffc1563c198 error 4 in ld-musl-x86_64.so.1[148576502000+53000] likely on CPU 12 (core 4, socket 0)

 

Test Plex without hardware encoding enabled.

 

See also https://docs.unraid.net/unraid-os/release-notes/6.12.0 about known issues with i915 driver

Using Intel 10th gen CPU so don't think it's related 

 

I can't actually turn the hardware transcode off I have a LOT of family and friends on it almost all the time. Without hardware transcoding I can handle 3-4 streams. 

 

With hardware I have had 40 with 25 transcoding at once. 

 

 

Since removing modprobr from go file and switching to the spinner cache pool things have been much more stable

 

 

One outage that corrected itself less than 2 min later.