(2014-02-13, 11:42)Jan0sh Wrote: I will retry this afternoon. Yesterday when the error occurred the partition where the files are on was completely full, no free byte left.
I will fill it up again and retest. Strange
So finaly found time for the test.
To fill up the HDD i used
Code:
dd if=/dev/zero of=largeFile bs=8G count=1
after a while the skipping sound error reoccured (the hdd is not completely full)
If I redo the dd command multiple times short behind each other, the movie will require to buffer again, but then even if no IO-ops is performed the movie will play without sound and very slow/juddery (frame by frame). Looks like the buffer is never filled up again (even it states before 1 to 100% buffered). One has to stop the current playback and restart it to get it working again.
It is totaly possible that there was some IO when the problem occured first, some unraring or so, but I would have thought, thosw would be of lower priority than the xbmc process. The dd command puts quit a lot load on CPU so it is also possible the problem is more cpu-stress related than IO related?.
The HDDs are quit fast.
Code:
dd if=/dev/zero of=large4 bs=8G count=1
2147479552 Bytes (2,1 GB) kopiert, 23,1502 s, 92,8 MB/s
I assume I can live with those skips-on-heavy-io but would be interested in why the system does not priorities xbmc instead of some random Io generating subprocess?
Also the issue with rebuffering seems to require some attention. Even if xbmc has to rebuffer (because of heavy io) the movie should play back just fine after the io is finished.
If there is anything I could test in addition just say it
Edit:
Did test if it could be an cpu issue with stress -cpu 2 . no problems occured, must have been io