[mythtv] mythweb stalls mythbackend, remote slave (all running .16 release)

[David Shay]

I have had my mythbackend "stall" on multiple occasions this week, and
believe I have narrowed the problem down to mythweb.  I have a slave backend
which does *not* share a filesystem with my master frontend/backend.  I just
recently started recording shows on this slave backend.  Twice this week, I
have gone to check on my shows and found mythbackend "stalled".  Basically,
it was still running, but not recording anything. If you tried to run
"mythbackend --printsched" , I get this:

2004-09-24 00:13:04 ReadStringList timeout (quick).
2004-09-24 00:13:04 Unexpected response to MYTH_PROTO_VERSION:

If I restart mythbackend, everything proceeds as expected.

If, however, I go into mythweb to the "Recorded Programs" screen, I get the
list of recorded programs.  For the one program that was recorded on the
slave backend, I do not get a preview image.  I suppose this is
understandable, given the fact that I am not NFS sharing these directories.
I don't really want to have all recordings go to a single shared directory,
but rather use all available space on both units.  The real problem, though,
is that the mere act of going to the recorded programs screen in this case
stalls mythbackend.  I get the following error in my mythbackend log file:

NULL in string in list in WriteStringList

After this, mythbackend is still running, but unresponsive (see the message
above for the errors that mythbackend --printsched gives).

I don't particularly care about not having a preview image, but I'd really
like to not have mythweb actually crash/stall the backend.

Any ideas?  Any more info I can provide?