The bottle neck I'm betting is the disk IO on the rrd files. Take a look at using better disks/disk setup, storing the rrd files seperate from the OS drive and swap. Raid 1, 0 or 10 will help a lot.
________________________________ From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Jon Greaves Sent: Wednesday, September 12, 2007 3:18 AM To: [email protected] Subject: [smokeping-users] Smokeping performance tracking Hi Tobi, Just an idea for tracking how smokeping is performing could you get smokeping to graph the time each probe takes to run. That way you should get a good indication if you are pushing a single probe to hard. We have does some basic testing our setup is with 130 groups and around 950 devices with the following results Setup: 20 ping every 5 mins 1 probe Result: the machine struggled to finish 1 full probe cycle under 300 seconds Setup 20 pings every 5 mins 1 probe per group (130 probes in total) Result: waitio time hit 100% and eventually the box crashed Setup 10 pings every 5 mins 1 probe per group (130 probes in total) Result: waitio time hit 100% and eventually the box crashed but took longer than above Setup 10 pings every 5 mins 10 probes split across the 130 groups Result: stable but not sure how much overhead we have (see above idea for tracking) We are hoping to push the setup up to 2000+ devices monitored once we get to that I'll post back some additional results. Current hardware spec OptiPlex GX520 Intel 1 * P4 HT 3.4gHz 2Gb RAM JonG The information contained in this e-mail and its attachments is confidential. It is intended only for the named address(es) and may not be disclosed to anyone else without Attenda's consent.
_______________________________________________ smokeping-users mailing list [email protected] https://lists.oetiker.ch/cgi-bin/listinfo/smokeping-users
