[mythtv-users] RE:FW: Potentially Useful Scripts

[William Powers]

I've had a variation of the first script running for quite a while with 
useful results.  I did not have a 0B file checker, though, and since I 
lost three movies last night to that particular annoyance, I think it's 
time I added that to my 'mythmon' script.

I have the following thoughts for your consideration:

1)  Unloading ivtv with the backend still running is problematic, since 
the module is in use.  I suggest stopping the backend, unloading the 
modules, (re)loading the modules, and (re)starting the backend.

2)  I have two pvr-250's and I've had the second encoder start to 
produce 0B files while the first encoder was in the middle of recording 
a valid (?) program.  For that reason, I'm considering adding an 
additional check so that the 0B recovery steps don't execute unless the 
number of 0B files with a current timestamp is greater than zero AND is 
equal to the total number of files with a current timestamp, like so:

NOW=`date '+%b %e %R'`
RUNNING=`eval ls -l $VIDEODIR/*\.nuv | grep "$NOW" | wc -l | awk '{ 
print $1 }'`
ZERO=`eval ls -l $VIDEODIR/*\.nuv | grep " 0 $NOW" | wc -l | awk '{ 
print $1 }'`

if [ $ZERO -gt "0" ]; then
	if [ $ZERO -eq $RUNNING ]; then
		/etc/rc.d/init.d/$PROCESS stop
		rmmod -r ivtv tuner msp3400 saa7115
		modprobe ivtv
		/etc/rc.d/init.d/$PROCESS start
		/bin/date +"Mythbackend restarted (0B file) at: %c" >> $LOG
		exit 1
	fi
fi

This is, of course, based on the assumption that the first recording in 
progress is not impacted at all by the driver hiccup, which is by no 
means guaranteed.  If both recordings always turn out to be toast, then 
it would obviously be better to unload the drivers immediately once the 
first 0B file shows up.

For my part, I can run using only a single encoder for days on end 
without incident, but as soon as encoder 2 kicks in, something goes 
wrong more than half of the time.  Until recently, I was stuck at 
myth-0.11 and couldn't upgrade for some screwy reason and with that 
version I would usually get a backend crash.  Last week I did a clean 
install of fedora core 1 & atrpms myth, ivtv, et al, and now that 
everything is current, I haven't had a backend crash (yet - fingers 
crossed).

But now the ivtv drivers seem to hiccup up whenever one encoder is 
running and the other one starts a new recording.  Maybe this will be 
helped by the (pending) new version of the ivtv driver.

Bill