On 26-Jun-05, at 7:27 PM, Isaac Richards wrote:
The id child's possibly the leaking element.
I added ~ScheduledRecording() to delete that and also m_pginfo but it
doesn't seem to have helped. I also applied the change in ticket 18 but
it didn't help either. I mean, they're legitimate leaks but they're not
bringing me to stability. My searching will continue as weather
permits... :)
If anyone is interested in playing along at home I've appended the
monitoring script I use. It gives a running count of loops and the
amount of xml data that is coming back and when the server's vsz
changes it prints the new size, how long it took to get there and the
number of bytes of the change.
Really shouldn't be polling the
server, though, since it _will_ send out messages on schedule changes
if
asked to.
I don't see another way to display current server status. My screen
shows what's recording (if anything), what jobs are active (with
progress percentage) and when the server is going to record again. I
also monitor the logfile to pick up "shutting down in NN seconds"
messages. I thought about going to the database to get this stuff from
the logging table but it was too ugly.
- George
#!/bin/bash
MythPid=$(ps --no-headers -C mythbackend | cut -c1-5)
echo Backend is pid $MythPid
LastSize=0 Loops=0
while :; do
MythSize=$(ps --no-headers -p $MythPid -o vsz)
if [[ $MythSize != $LastSize ]]; then
printf "\r$MythSize after $Loops loops +
$((MythSize-LastSize))\n"
LastSize=$MythSize
Loops=0
fi
XmlSize=$(wget --quiet -O - http://127.0.0.1:6544/xml | wc -c)
printf "\r$((++Loops)) ($XmlSize)"
done
_______________________________________________
mythtv-dev mailing list
[email protected]
http://mythtv.org/cgi-bin/mailman/listinfo/mythtv-dev
