analyzing oVirt infra performance and bottle necks
fyi, thanks to kiril we now how cacti statistics on all oVirt servers, which should help us analyse and find the chock points that mostly Jenkins slaves suffer from, so we can more accurately decide which servers do we need to get for the new infra. please spend a few moments before tomorrow infra meeting and review the current status [1] so we can discuss it on the meeting. credentials for access are on the usual place. Eyal. [1] http://monitoring.ovirt.org/cacti/graph_view.php?action=tree&tree_id=1&leaf_id=25
On Sun, 5 Jan 2014, Eyal Edri wrote: Drilling through the graphs I note the following: 1. SNMP connection problems to: foreman.ovirt.org and artifactory.ovirt.org 2. The jenkins buildslaves in VMs are heavily loaded for CPU 3. the rackspace uinits (01 and 02) are heavily loaded for CPU 4. no disk stats are displayed on rackspace 01 or 02 but drilling on rackspace02.ovirt.org - Disk IO Bytes - sdb is pretty grim -- Russ herrold
I would like somebody to take a look on missing graphs issues. I tried to take a look but no success. Any help will be appreciated. Kiril ----- Original Message -----
From: "R P Herrold" <herrold@owlriver.com> To: "oVirt infrastructure ML" <infra@ovirt.org> Sent: Monday, January 6, 2014 4:33:35 PM Subject: analyzing oVirt infra performance and bottle necks
On Sun, 5 Jan 2014, Eyal Edri wrote:
Drilling through the graphs I note the following:
1. SNMP connection problems to: foreman.ovirt.org and artifactory.ovirt.org
2. The jenkins buildslaves in VMs are heavily loaded for CPU
3. the rackspace uinits (01 and 02) are heavily loaded for CPU
4. no disk stats are displayed on rackspace 01 or 02 but drilling on rackspace02.ovirt.org - Disk IO Bytes - sdb is pretty grim
-- Russ herrold _______________________________________________ Infra mailing list Infra@ovirt.org http://lists.ovirt.org/mailman/listinfo/infra
participants (3)
-
Eyal Edri -
Kiril Nesenko -
R P Herrold