2016-06-17 10:03 GMT+02:00 Christian Balzer <chibi@xxxxxxx>: > I'm unfamilar with Xen and Xenserver (the later doesn't support RBD, btw), > but if you can see all the combined activity of your VMs on your HW in the > dom0 like with KVM/qemu, a simple "iostat" or "iostat -x" will give you the > average IOPS of a device. > Same of course within a VM. I'm able to see the combined activity directly from Dom0. With "iostat" should I look for 'tps' column ? > However that's the average, you're likely to have peaks much higher than > that. > For this you'll either have to collect and graph that data > (collectd/graphite, etc) and/or run something like atop during peak hours > and watch it or have it write logs with a high sample rate. > As in, atop can keep a log of all states, but the default interval of 10 > minutes with Debian is likely too course to spot real peaks. > See the atop documentation. Running "iostat" every seconds, i can see about 800-1000 tps (transactions per seconds) I can try to install "atop" but i'm totally new to this, I don't know how use. Any hint? Currently i'm reading the "atop" man page but i don't want to run in for the whole weekend with wrong parameters. _______________________________________________ ceph-users mailing list ceph-users@xxxxxxxxxxxxxx http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
 |