On 12/23/2014 01:36 PM, Robert van Leeuwen wrote: >> from above numbers it doesn't sound like storage node is the >> culprit - 50% drop happens on compute node going from baremetal to >> virtual. So I'm inclined to think it's a tuning of virtio (if that >> is even possible). > > Some tuning tips for kvm: http://www.linux-kvm.org/page/Tuning_KVM > > We saw a major improvement using RAW vs QCOW2 (in theory this should > be similar, however, when you need a high random IO throughput we did > see major improvements in our setup with RAW) > > On more thing to check is partition alignment. (not very likely for a > recent OS)
thank you Robert. As it turned out - it was a bug in our nova.conf
template which specified virt_type as qemu instead of kvm.. I found that
out comparing line-by-line manually created VM XML on the same node
(that performed great) and the one generated by nova. d'oh!
--
Dmitry Makovey
Web Systems Administrator
Athabasca University
(780) 675-6245
---
Confidence is what you have before you understand the problem
Woody Allen
When in trouble when in doubt run in circles scream and shout
http://www.wordwizard.com/phpbb3/viewtopic.php?f=16&t=19330
signature.asc
Description: OpenPGP digital signature
_______________________________________________ Mailing list: http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack Post to : [email protected] Unsubscribe : http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack
