-
Bug
-
Resolution: Fixed
-
Major
-
None
-
Flounder
The --gst-record feature in Weston rarely starts successfully in my testing, it usually fails with this in weston.log:
[16:50:27.188] Output virtual1, ()
mode 1920x1080@60.0
[16:50:27.189] gst_recorder_create (1920x1080) crop 384x368 at 0,0
[16:50:27.189] gst_pipeline: starting: appsrc name=src ! omxh264enc target-bitrate=300000 control-rate=2 no-copy=true name=my_encoder ! video/x-h264,width=384,height=368 ! rtph264pay config-interval=1 name=my_h264pay ! queue ! udpsink sk
[16:50:27.219] allocation query failed
[16:50:27.219] failed to find OMX buffer pool
In my local testing, it works perhaps 1 time in 10 attempts. It then crashes when mapviewer attempts to use the virtual output, but I've opened a separate issue for that. At the moment, this problem makes it extremely difficult to test the cluster setup, and I doubt we could demo at CES with it like this. So either it needs to be fixed, or a judgement call made about whether to try backporting the Waltham transmitter work. I'm going to poke around in the gst-recorder code some to see if I can make any headway, but realistically someone at Renesas or Cogent need to look at this.