On 2/9/2005 8:09 PM, Justin Walther wrote:
I too was seeing this exact same problem. I was using ext3 as my FS
on my recordings partition. I changed my filesystem to XFS and I no
longer get the IOBOUND errors.
HOWEVER, I am now once again experiencing the recordings ending early.
It always happens when its trying to record 2 programs at once and
the backend dies as soon as it autoexpires 1 or 2 programs to make
room for the new recordings. This happens usually 1-3 mins into the
recordings. This all makes me believe its still a throughput issue
but since I am no longer seeing the IOBOUND errors, I'm stumped.
I AM using a LVM volume comprised of 2 harddrives so perhaps the
performace of the LVM volume can't handle the traffic of recording 2
programs and expiring (deleting) 2 programs. Still...it seems ugly
that this would actually kill the backend.
Sorry I can't help with the troubles you are experiencing, but I can
tell you it is highly doubtful it is LVM. I am running XFS on LVM
(using 4 drives though). With two capture cards, a slave backend with
two HDTV capture cards, and a remote frontend. I can record from all 4
cards at once, while it is doing commercial flagging (for a program that
was on the previous hour), and I am watching a previously recorded
program. The slave backend writes the HDTV stream to the master backend
via NFS. I don't auto-expire, but other than that I hammer the snot out
of this thing.
I wish we could get some answers from someone more knowledgable. :)
Sorry, maybe someone else could chime in?
--
David
_______________________________________________
mythtv-users mailing list
[email protected]
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-users
