Server Version#: 1.25.2.5319
Fedora Linux 35 with kernel 5.15.12-200.fc35.x86_64
CPU i5-12600K with Intel Alderlake_s (Gen12) graphics.
When hardware transcoding is enabled Plex Transcoder crashes every now and then taking Plex server with it. dmesg shows:
[ 5926.025586] i915 0000:00:02.0: [drm] Resetting vcs1 for preemption time out
[ 5926.033686] i915 0000:00:02.0: [drm] GPU HANG: ecode 12:4:28fffffd, in Plex Transcoder [285442]
[ 5929.934630] i915 0000:00:02.0: [drm] GPU HANG: ecode 12:4:28fffffd, in Plex Transcoder [285442]
[ 5929.937372] i915 0000:00:02.0: [drm] Resetting vcs1 for stopped heartbeat on vcs1
[ 5929.938769] i915 0000:00:02.0: [drm] Resetting chip for stopped heartbeat on vcs1
[ 5930.041926] [drm:__uc_sanitize [i915]] *ERROR* Failed to reset GuC, ret = -110
[ 5930.115066] i915 0000:00:02.0: [drm] *ERROR* Failed to reset chip
[ 5930.117918] i915 0000:00:02.0: [drm:add_taint_for_CI [i915]] CI tainted:0x9 by intel_gt_reset+0x2e0/0x300 [i915]
[ 5930.220925] [drm:__uc_sanitize [i915]] *ERROR* Failed to reset GuC, ret = -110
[ 5930.224651] i915 0000:00:02.0: [drm] Plex Transcoder[285442] context reset due to GPU hang
[ 5930.305555] show_signal_msg: 9 callbacks suppressed
[ 5930.305557] Plex Media Serv[285776]: segfault at 0 ip 0000000000000000 sp 00007f5871239788 error 14
[ 5930.306688] Code: Unable to access opcode bytes at RIP 0xffffffffffffffd6.
Since it starts with GPU hang I’d think the cause is with the CPU model, but I’ve manually encoded and transcoded plenty of movies using ffmpeg without any issue. So seems to be something specific to the Plex build that triggers the hang.
Has anyone seen this or any idea what the issue might be?