Hello Adrian,
On 13 Jun 2013, at 15:32, Adrian Sevcenco wrote:
> On 06/09/2013 11:26 AM, Sergio Ballestrero wrote:
>> We run a single gmetad with 11 collector gmond (listen-only) and 1
>> "client" gmond (send-only) to monitor the host itself, all using unicast.
> Hi! Are this gmonds on the same (gmetad) host? i imagine that yes, so i
> ask you : how did you specify the each configuration?
> was enough to replace in /etc/init.d/gmond
> GMOND=/usr/sbin/gmond
> with
> GMOND=/usr/sbin/gmond -c /etc/ganglia/my_conf_1
> ?
It was not really enough, because then the service stop kills all gmond
indiscriminately.
I've copied the "server" script at bottom, feel free to use it. Then the
"client" gmond is the same except for replacing where needed
for CFG in /etc/ganglia/gmond.conf; do
I keep them separate for ease of management (e.g. restart all client gmond)
> Each gmond collector defines a separate cluster?
Yes, grouped by function. The majority of them is actually all in one cluster,
the HLT worker nodes.
> For this number of nodes how did you chose the size of rrd storage? (the
> time granularity) how big do you have the rrd archive?
For the moment we are using the defaults, so about 500KB per node, 1.3GB total
(probably some duplication from moving hosts across groups)
We'll probably want to tune it later, try to keep high res for a bit longer to
help debugging performance issues older than one hour.
Cheers,
Sergio
-------------- sysV Init script -------------
[sash@pc-tdq-sys-02 ganglia]$ cat /etc/init.d/gmondsrv
#!/bin/sh
#
# chkconfig: - 70 40
# description: gmond server startup script
#
GMOND=/usr/sbin/gmond
. /etc/rc.d/init.d/functions
RETVAL=0
case "$1" in
start)
echo -n "Starting GANGLIA gmond: "
[ -f $GMOND ] || exit 1
RETVAL=0
for CFG in /etc/ganglia/gmond.*.srvconf; do
PID=/var/run/$(basename $CFG).pid
daemon --pidfile $PID $GMOND -c $CFG -p $PID
RETVAL=$[RETVAL||$?]
done
[ "$RETVAL" -eq 0 ] && success $"$base startup" || failure $"$base
startup"
echo
[ $RETVAL -eq 0 ] && touch /var/lock/subsys/gmond
;;
stop)
echo -n "Shutting down GANGLIA gmond: "
RETVAL=0
for CFG in /etc/ganglia/gmond.*.srvconf; do
PID=/var/run/$(basename $CFG).pid
killproc -p $PID gmond
R=$?
RETVAL=$[RETVAL||R]
done
[ "$RETVAL" -eq 0 ] && success $"$base startup" || failure $"$base
startup"
echo
[ $RETVAL -eq 0 ] && rm -f /var/lock/subsys/gmond
;;
restart|reload)
$0 stop
$0 start
RETVAL=$?
;;
status)
RETVAL=0
for CFG in /etc/ganglia/gmond.*.srvconf; do
PID=/var/run/$(basename $CFG).pid
status -p $PID gmond
RETVAL=$[RETVAL||$?]
done
;;
*)
echo "Usage: $0 {start|stop|restart|status}"
exit 1
;;
esac
exit $RETVAL
------------------------------------------------------------------------------
This SF.net email is sponsored by Windows:
Build for Windows Store.
http://p.sf.net/sfu/windows-dev2dev
_______________________________________________
Ganglia-general mailing list
Ganglia-general@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/ganglia-general
