[mythtv-commits] Ticket #12602: lossless transcode that fills the disk loses the tail of the recording

[MythTV]

#12602: lossless transcode that fills the disk loses the tail of the recording
------------------------------------------+---------------------------
 Reporter:  Timothy D Witham <twitham@…>  |          Owner:  jyavenard
     Type:  Bug Report - General          |         Status:  new
 Priority:  minor                         |      Milestone:  unknown
Component:  MythTV - Mythtranscode        |        Version:  0.27.5
 Severity:  medium                        |     Resolution:
 Keywords:  disk full transcode truncate  |  Ticket locked:  0
------------------------------------------+---------------------------

Comment (by Timothy D Witham <twitham@…>):

 Looks like mythtv code is at a higher abstraction layer that is not
 actually doing the open/write/close calls at all.  transcode.cpp calls
 avfw->CloseFile(); which in avformatwriter.cpp does an avio_close without
 checking its return value:
 https://code.mythtv.org/doxygen/avformatwriter_8cpp_source.html#l00211

 avio_close is over in ffmpeg's libav*:
 https://ffmpeg.org/doxygen/trunk/aviobuf_8c_source.html#l00943

 This calls an avio_flush which hopefully blocks until the data is written,
 then avio_close returns ffurl_close which returns h->prot->url_close:
 https://ffmpeg.org/doxygen/trunk/avio_8c_source.html#l00390

 I still haven't tracked down the actual close() in libav, but so far all
 calls are passing the return value back up the chain, until the return of
 avio_close is ignored in mythtv's avformatwriter.cpp.

 So my bug fix idea should still work only if libav detects the ENOSPC
 error and passes it back up to avio_close where it can be detected and
 handled better by mythtv code (log the error and remove truncated new file
 instead of rename over original).

--
Ticket URL: <https://code.mythtv.org/trac/ticket/12602#comment:4>
MythTV <http://www.mythtv.org>
MythTV Media Center