PMS freezes entire host system (Ubuntu/Linux) when HW transcoding MPEG2 on Intel Skylake

Server Version#: 1.41.7.9823
Player Version#: 10.28.0.3441 (Onn 4k Pro client)

Hi all, I am consistently experiencing complete host server freezes while PMS is transcoding MPEG2 files recorded from a HDHomerun Connect Duo. I first noticed this when transcoding live TV streams, but then tried transcoding a previously recorded show which ended with the same result, complete system freeze after around 5 or 10 mins of transcoding. This seems to be limited to MPEG2 transcodes as there are no issues with HW transcoding of any other video file formats such as H264, HEVC, etc, As far as I can tell this does not happen with software transcoding either. I have read about issues with Skylake and hyperthreading and I had updated to the latest microcode years ago which should fix that issue. I have also disable hyperthreading for good measure, but that has not resolved the issue. This has been ongoing for several months, maybe longer, with previous versions of PMS.

I am running PMS on an Intel NUC6 with Core i3-6100U (Skylake) processor (Intel HD 520) behind Ubuntu 24.04 using the i915 driver. This is the output from vainfo:

brandon@brandon-desktop:~$ sudo vainfo
error: XDG_RUNTIME_DIR is invalid or not set in the environment.
error: can’t connect to X server!
libva info: VA-API version 1.20.0
libva info: Trying to open /usr/lib/x86_64-linux-gnu/dri/iHD_drv_video.so
libva info: Found init function __vaDriverInit_1_20
libva info: va_openDriver() returns 0
vainfo: VA-API version: 1.20 (libva 2.12.0)
vainfo: Driver version: Intel iHD driver for Intel(R) Gen Graphics - 24.1.0 ()
vainfo: Supported profile and entrypoints
VAProfileNone : VAEntrypointVideoProc
VAProfileNone : VAEntrypointStats
VAProfileMPEG2Simple : VAEntrypointVLD
VAProfileMPEG2Simple : VAEntrypointEncSlice
VAProfileMPEG2Main : VAEntrypointVLD
VAProfileMPEG2Main : VAEntrypointEncSlice
VAProfileH264Main : VAEntrypointVLD
VAProfileH264Main : VAEntrypointEncSlice
VAProfileH264Main : VAEntrypointFEI
VAProfileH264Main : VAEntrypointEncSliceLP
VAProfileH264High : VAEntrypointVLD
VAProfileH264High : VAEntrypointEncSlice
VAProfileH264High : VAEntrypointFEI
VAProfileH264High : VAEntrypointEncSliceLP
VAProfileVC1Simple : VAEntrypointVLD
VAProfileVC1Main : VAEntrypointVLD
VAProfileVC1Advanced : VAEntrypointVLD
VAProfileJPEGBaseline : VAEntrypointVLD
VAProfileJPEGBaseline : VAEntrypointEncPicture
VAProfileH264ConstrainedBaseline: VAEntrypointVLD
VAProfileH264ConstrainedBaseline: VAEntrypointEncSlice
VAProfileH264ConstrainedBaseline: VAEntrypointFEI
VAProfileH264ConstrainedBaseline: VAEntrypointEncSliceLP
VAProfileVP8Version0_3 : VAEntrypointVLD
VAProfileVP8Version0_3 : VAEntrypointEncSlice
VAProfileHEVCMain : VAEntrypointVLD
VAProfileHEVCMain : VAEntrypointEncSlice
VAProfileHEVCMain : VAEntrypointFEI

Here is the Plex Media Server Log (crash at 14:14 at the end of the log file). Seems the crash was sudden enough that nothing helpful was logged.

I tried deleting all the codecs in the /var/lib/plexmediaserver/Library/Application Support/Plex Media Server/Codecs directory which also did not resolve the issue.

I have performed a lengthy memory test and the system RAM checks out. I have also done some CPU stress tests with no issues found. I have also replaced the system power supply to ensure that was not the issue. I don’t know where to go from here to diagnose this issue.

Thank you for the logs.

  1. Your server is actually transcoding and going into sloth mode (buffers are full and waiting to send to player)

  2. Player is having connectivity problems.

Jul 15, 2025 13:46:33.797 [136405989030712] DEBUG - Content-Length is 10197893 (of total: 10197893).
Jul 15, 2025 13:46:33.816 [136406288235320] DEBUG - CERT: incomplete TLS handshake from [::ffff:192.168.1.243]:46874: sslv3 alert certificate unknown (SSL routines)
Jul 15, 2025 13:46:33.818 [136406288235320] DEBUG - CERT: incomplete TLS handshake from [::ffff:192.168.1.243]:46878: sslv3 alert certificate unknown (SSL routines)
Jul 15, 2025 13:46:33.889 [136406288235320] DEBUG - CERT: incomplete TLS handshake from [::ffff:192.168.1.243]:46904: sslv3 alert certificate unknown (SSL routines)
Jul 15, 2025 13:46:33.908 [136405938629432] DEBUG - Request: [192.168.1.243:46894 (Subnet)] GET /:/timeline?

I do see the crash / panic as transcoding ends.

Jul 15, 2025 14:14:35.507 [136406290344760] DEBUG - Removed transcode data consumer, active count 1 => 0
Jul 15, 2025 14:14:35.863 [136405938629432] DEBUG - [Req#ffc1c/Transcode/295772c6ccd58c7a-com-plexapp-android/bd8fe70e-0ec4-4b7b-856a-b3d9d1a4d7db] Session 295772c6ccd58c7a-com-plexapp-android (3) is unthrottling
Jul 15, 2025 14:14:35.864 [136405894273848] DEBUG - [Req#ffc1f/Transcode/295772c6ccd58c7a-com-plexapp-android/bd8fe70e-0ec4-4b7b-856a-b3d9d1a4d7db] Throttle - Getting back to work.
��������������������������������������������������������������

Notice PMS already had finished playback.

I recommend two steps for diagnosing further.

  1. PMS 1.41.7 had a known database bloat problem.
    I recommend upgrading to PMS 1.41.8 as soon as possible.

  2. Turn off IPv6 (Settings - server - Network - Show Advanced) because you don’t have IPv6 and you’re getting strange IPv6 SSLv3 certificate error messages from the Android.

Thanks for the assistance Chuck. I have updated PMS to version 1.41.9.9961 and disabled IPv6 as you recommended. The “incomplete TLS handshake” messages still persist, but without the IPv6 address stuff. Unfortunately, the system crashes have also persisted. After more testing, I have found that this is occurring with HW transcoding of HEVC source files as well, but I have only reproduced that when streaming to my Onn 4k Pro, it did not happen when streaming the same HEVC source file transcoded (HW) to my Windows client after over 1.5 hours of transcoding. However, when streaming transcoded MPEG2 to the Windows client, it caused a complete system crash within 4 mins. It may not be limited to MPEG2, but it certainly seems more prevalent with MPEG2 files. Here is of a summary of the testing I have done:

Client - Onn 4k Pro

  1. MPEG2 source transcoded (HW) - 8 mbit 1080p → crash occurred within 5 minutes
  2. HEVC source transcoded (HW) - 8 mbit 1080p → crash occurred after about 30 mins

Client - Windows PC with Plex from Microsoft Store

  1. HEVC source (same as above) transcoded (HW) - 8 mbit 1080p → 1.5 hours, no crash
    [No log]
  2. MPEG2 source (same as above) transcoded (HW) - 8 mbit 1080p → crash in ~4 mins.

I have noticed that there are none of those “incomplete TLS handshake” messages when streaming to the Windows client.

I monitored CPU and memory (12 GB) usage and temperature during all of this testing. No issues there.

Thanks for the logs.

Please remove VaapiDriver="i965" from your Preferences.xml.

It’s no longer applicable and can actually crash the host.

-vsync cfr -init_hw_device vaapi=vaapi:/dev/dri/renderD128,driver=i965 -
1 Like

I do not see VaapiDriver or i965 anywhere in my Preferences.xml. Is there anything else that might be causing it to use the i965 driver? Can/should I add VaapiDriver=“iHD” to the Preferences.xml?

I am also noticing there is a symlink to iHD_drv_video.so in the PMS Cache/va-dri-linux-x86_64/ folder but the actual file seems to be missing from the plex Drivers folder.

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.