[mythtv-users] 0.21 and database corruption/how to rebuild and update from backup
Neil Bird
neil at fnxweb.com
Tue Mar 18 08:38:51 UTC 2008
Around about 18/03/08 05:27, Bob Sully typed ...
> One minor concern... The backend segfaulted, and I had to restart
> it manually. Hopefully that won't happen again. I'm wondering if maybe I
> should write a little script (if there isn't one already out there) to
> check every so often whethere the backend is running, and if not, restart
> it...?
As I'm paranoid, and it does stuff up sometimes, I have (among other
entries for mythfilldatabase and LCD status):
[/etc/cron.d/mythtv]
SHELL=/bin/bash
PATH=/sbin:/bin:/usr/sbin:/usr/bin:/usr/local/bin
MAILTO=root at mymainbox
HOME=/
*/5 * * * * root /usr/local/bin/check-mythbackend
And the script (now this may look a little odd but I figured, for no good
reason, really) that 12 cron runs an hour with 5 checks apiece was better
that 60 runs an hour:
[/usr/local/bin/check-mythbackend but edited to remove some other stuff]
#!/bin/bash
shopt -s nullglob
doit=false
# Need to start mythbackend?
if [ -e /var/run/mythbackend.pid ]; then
pid=`cat /var/run/mythbackend.pid`
if [ ! -e /proc/$pid ]; then
rm -f /var/run/mythbackend.pid
doit=true
fi
fi
$doit && /sbin/service mythbackend start 2>&1 >/dev/null &
doit=false
# Check 4 more times
for attempt in 1 2 3 4; do
sleep 60
doit=false
# Need to start mythbackend?
if [ -e /var/run/mythbackend.pid ]; then
pid=`cat /var/run/mythbackend.pid`
if [ ! -e /proc/$pid ]; then
rm -f /var/run/mythbackend.pid
doit=true
fi
fi
$doit && /sbin/service mythbackend start 2>&1 >/dev/null &
done
YMMV.
--
[neil at fnx ~]# rm -f .signature
[neil at fnx ~]# ls -l .signature
ls: .signature: No such file or directory
[neil at fnx ~]# exit
More information about the mythtv-users
mailing list