[QUESTION] kvssink is taking 2 minutes to shut down after EOS is sent to the Gstreamer pipeline

[bkovacs-orca]

I have a Gstreamer pipeline that ends with kvssink, and sometimes it is taking 2 minutes to shut down after EOS is sent to the Gstreamer pipeline. In what situations is this possible? What can be the reason for this?

** Logging ** I don't have logs yet, because I don't have an easy way to reproduce this case, it happens sporadically.

** Any design considerations/constraints ** Using the following Gstreamer pipeline:

appsrc name=appsrcvideo block=false leaky-type=downstream max-bytes=5000000 ! h264parse name=parser ! kvssink name=kvssink stream-name=X storage-size=512 streaming-type=2 file-start-time=0 retention-period=1 access-key=X secret-key=X aws-region=us-west-2 stream-tags=X  appsrc name=appsrcaudio block=false leaky-type=downstream max-bytes=5000000 ! audio/mpeg,mpegversion=4,rate=(int)48000,channels=1,stream-format=(string)raw,codec_data=(buffer)1188 ! aacparse ! queue ! kvssink.