-
Notifications
You must be signed in to change notification settings - Fork 296
Description
Observed Behavior
CasparCG runs normally for some time, but when multiple SRT inputs are active simultaneously, the server process terminates unexpectedly and playback stops.
No application-level CasparCG logs are produced before the crash.
Systemd reports the following at the time of the failure:
Jan 22 11:22:26 : casparcg.service: Main process exited, code=dumped, status=4/ILL
Jan 22 11:22:26 : Failed with result 'core-dump'.
Jan 22 11:22:26 : Consumed 3h 11min 22.359s CPU time.
Expected behaviour
CasparCG should continue running while decoding multiple SRT inputs and routing them into the GRID mixer.
If a codec or consumer configuration is not supported, the server should fail gracefully without crashing.
Steps to reproduce
1.MIXER 19 GRID 2 0 Linear
PLAY 1-10 "srt://" -f mpegts -fflags +genpts+discardcorrupt -err_detect ignore_err -analyzeduration 300000 -probesize 3000000
PLAY 19-2 "srt://" -f mpegts -fflags +genpts+discardcorrupt -err_detect ignore_err -analyzeduration 300000 -probesize 3000000
PLAY 19-3 "srt://" -f mpegts -fflags +genpts+discardcorrupt -err_detect ignore_err -analyzeduration 300000 -probesize 3000000
PLAY 19-4 "srt://" -f mpegts -fflags +genpts+discardcorrupt -err_detect ignore_err -analyzeduration 300000 -probesize 3000000
PLAY 19-1 ROUTE://1-10 CUT 0
PLAY 19-2 ROUTE://2-10 CUT 0
PLAY 19-3 ROUTE://3-10 CUT 0
PLAY 19-4 ROUTE://4-10 CUT 0
MIXER 19-1 VOLUME 0
MIXER 19-2 VOLUME 0
MIXER 19-3 VOLUME 0
MIXER 19-4 VOLUME 0
2.
3. ...
Environment
CasparCG Server: 2.5.0 Dev (dca9b6e)
OS: Ubuntu 24.04 (AWS EC2)
Execution: systemd service
Headless server (TTY only, no X11, no Wayland)
CPU: Intel Xeon Platinum 8259CL (16 vCPU, x86_64)
Memory: 62 GB RAM
GPU: No hardware GPU available (CPU-only instance)