search

teltek / Galicaster

The Galicaster Project is an open initiative to provide flexible, state-of-the-art solutions for recording educational multimedia contents like lectures and conferences
galicaster.teltek.es
Other
35 stars 31 forks source link

GC 2.x: Timeout trying to receive EOS message #425

Closed smarquard closed 7 years ago

smarquard commented 8 years ago

On one CA with an IP camera we have intermittently got an error when the recording stops when it appears that the gstreamer pipeline cannot be shut down properly. We're not sure of the cause.

The following is logged to the console (stdout or stderr):

Traceback (most recent call last):
  File "/usr/share/galicaster/galicaster/scheduler/scheduler.py", line 122, in __stop_record
    self.recorder.stop()
  File "/usr/share/galicaster/galicaster/recorder/service.py", line 184, in stop
    self.__close_mp()
  File "/usr/share/galicaster/galicaster/recorder/service.py", line 192, in __close_mp
    self.current_mediapackage.status = mediapackage.RECORDED
AttributeError: 'NoneType' object has no attribute 'status'

The auto-recover then handles the mediapackage, which gets recovered with a "Recovered" prefix.

gstreamer logs:

2:18:46.087645621  4131      0x3034b70 WARN          v4l2bufferpool gstv4l2bufferpool.c:912:gst_v4l2_buffer_pool_stop:<gc-v4l2-src:pool:src> some buffers are still outstanding
2:19:12.421106747  4131 0x7fa968001850 WARN                 rtspsrc gstrtspsrc.c:2664:on_timeout:<gc-rtp-src> source 27867261, stream 27867261 in session 0 timed out

galicaster logs:

galicaster      2016-09-09 08:49:59,057 INFO    scheduler/scheduler     Timeout to stop record 56467754
galicaster      2016-09-09 08:49:59,077 INFO    recorder/service        Stopping the capture
galicaster      2016-09-09 08:49:59,077 DEBUG   recorder/recorder       Stopping recorder, sending EOS event to sources
galicaster      2016-09-09 08:50:29,277 ERROR   recorder/recorder       Timeout trying to receive EOS message
galicaster      2016-09-09 08:50:29,277 ERROR   recorder/service        Handle error (Timeout trying to receive EOS message)
galicaster      2016-09-09 08:50:29,642 INFO    mediapackage/repository Trying to recover the crashed recording
galicaster      2016-09-09 08:50:29,757 INFO    mediapackage/repository Copying file /usr/share/galicaster-repository/gc_menz10_20160718T09h55m24_37/org.opencastproject.capture.agent.properties to /usr/share/galicaster-repository/rectemp/org.opencastproject.capture.agent.properties
galicaster      2016-09-09 08:50:29,876 INFO    mediapackage/repository Repository temporal files moved to /usr/share/galicaster-repository/gc_menz10_20160718T09h55m24_37
galicaster      2016-09-09 08:50:30,490 INFO    mediapackage/repository Crashed recording added to the repository
galicaster      2016-09-09 08:50:30,491 INFO    mediapackage/repository Copying file /usr/share/galicaster-repository/gc_menz10_20160718T09h55m24_37/org.opencastproject.capture.agent.properties to /usr/share/galicaster-repository/gc_menz10_20160718T09h55m24_37/org.opencastproject.capture.agent.properties
galicaster      2016-09-09 08:50:30,537 DEBUG   recorder/service        Connecting recover recorder callback
galicaster      2016-09-09 08:50:30,642 INFO    opencast/service        Set status unknown to server

Also this but not sure if it's related (console / stdout or stderr)

Failed to connect to Mir: Failed to connect to server socket: No such file or directory
Unable to init server: Could not connect: Connection refused

We're not sure of the underlying cause, but it seems the Galicaster traceback should be fixed at least.

silviabf commented 7 years ago

The console error does not appears anymore. I close this issue. Thanks smarquard!