[rrd-users] Re: Directory names
stockrt at gmail.com
Mon Jul 24 18:06:59 MEST 2006
To centralize the graphing you can use SNMP to get the CPU usage in the
servers from one machine, using net-snmp package with snmpd on the servers.
Like, the management machine is the central, with net-snmp (snmp-walk,
snmp-get) and the the servers (monitored stations) wil run snmpd to answer
the snmp-get and snmp-walk requests.
This can be done through shell scripts, to collect data and store into local
rrd files, for graphing later. (sample in
You got it?
This is the good pratice, to use SNMP, but you can use some scripts to get
cpu and memory usage local from the servers, and then export this data to
remote rrd files (rrd files on the central management machine) via NFS,
remote SSH calls, or something else, but this is very plataform dependant
scripts and causes network overhead, not a good pratice.
Samples for linux here: http://stockrt.unicruz.edu.br/scripts/mrtg
On 7/24/06, Sanjeev Sharma <sanjeev.sharma at ideacellular.com> wrote:
We are very new to rrdtool. We are currently using mrtg
We have been trying to produce graphs for CPU and memory utilization for our
unix servers. Can we get some sample scripts to do that.
Our unix servers includes sun, HP-UX and linux and we want to install
rrdtool on only one sun server and we want to produce graphs for other
servers from there only.
Any help would be highly appreciated.
(55) 9985 2127
(55) 3332 5923
(55) 3333 8158
(55) 3321 1535
MSN: stockrt at hotmail.com
GTalk: stockrt at gmail.com
Unsubscribe mailto:rrd-users-request at list.ee.ethz.ch?subject=unsubscribe
Help mailto:rrd-users-request at list.ee.ethz.ch?subject=help
More information about the rrd-users