Numerous threads on transcode issues, server crashes, don’t use HW acceleration, …
Bottom line is PMS is unusable for live TV. I thought I had a work around to reinstall the video drivers but even that isn’t “fool proof”. I had it all working on AppleTV and then I switch to ROKU and the crashes began again. I turned off HW acceleration and the the ROKU still wouldn’t play LiveTV.
I know it’s convenient to blame Windows update and/or Intel drivers and/or sun spots but bottom line is PLEX is the only unstable thing on this machine.
Here’s the issue, if you want to charge for a product you need to make it usable.
Here’s my latest logs of crashes with HW acceleration off. PLEASE FIX ASAP and please don’t tell me how PLEX isn’t the problem…
As you would have seen there, these crashes which are within the intel driver mfx_mft_h264ve_32.dll are only happening for users running Plex Media Server as a service
I need confirmation that Hardware Acceleration is turned off. The logs suggest otherwise
May 12, 2018 20:12:35.359 [11628] DEBUG - Codecs: testing h264_qsv (encoder)
May 12, 2018 20:12:35.359 [11628] DEBUG - Codecs: hardware transcoding: testing API qsv
May 12, 2018 20:12:35.365 [11628] ERROR - [FFMPEG] - Failed to create Direct3D device
May 12, 2018 20:12:35.365 [11628] DEBUG - Codecs: hardware transcoding: opening hw device failed - probably not supported by this system, error: Unknown error occurred
Your logs are only as good as the messages placed in your code and are typically of little value in situations like this. PLEASE stop trying to blame things on someone else. Something changed in PLEX and things broke.
When do you plan to fix it. Something in the fact that the server keeps crashing indicates PLEX has issues. I’ll bet PLEX is passing a malformed on inappropriate argument to cause the crash. I tried 4 different versions going back almost 2 years of the video driver for my gen 4 i7. All behaved the same. If I pass a non-null terminated string to printf I’ll cause a crash. I’ll bet PLEX is the problem here.
Do you even do any testing when these things are reported? DOZENS of similar threads about these issues and all you seem to do if rely on trace files. Using an embedded trace and log files is a very inefficient way to debug. Great to get to the general problem but that’s about it.
As you would have seen there, these crashes which are within the intel driver mfx_mft_h264ve_32.dll are only happening for users running Plex Media Server as a service
I need confirmation that Hardware Acceleration is turned off. The logs suggest otherwise
May 12, 2018 20:12:35.359 [11628] DEBUG - Codecs: testing h264_qsv (encoder)
May 12, 2018 20:12:35.359 [11628] DEBUG - Codecs: hardware transcoding: testing API qsv
May 12, 2018 20:12:35.365 [11628] ERROR - [FFMPEG] - Failed to create Direct3D device
May 12, 2018 20:12:35.365 [11628] DEBUG - Codecs: hardware transcoding: opening hw device failed - probably not supported by this system, error: Unknown error occurred
@Smokindog said:
I’ve tried it both with acceleration on and off.
Yes - you did disable Hardware Acceleration at 20:13:23 on May 12th and the intel driver crashes were before that at 20:12:35 and 20:07:52
The intel driver crash has already been referred to the development team. Until there is a resolution to it, you have one of two choices to choose from
Do not run Plex Media Server as a service or as headless without a monitor and a windows desktop session, or
We are expecting these Intel h264_qsv encoder crashes to have been fixed by the ffmpeg update which has been released in Plex Media Server beta 1.17.0.1709-982421575
