Re: [Qemu-discuss] Disk Performance

[Quentin Hartman]

Thanks for the response. This performance problem is happening with disk io in the vm, how would network affect that? Even when I'm using ceph, all the networking is happening "behind" the vm, it just sees a disk. fwiw though, the initial qemu command line that was being used [1] does have virtio stuff specified for the network connection, but the problem still exists.

QH

1 - The full qemu command as generated by Openstack http://pastebin.com/ydNQpLqd

On Thu, May 29, 2014 at 8:38 AM, Dale R. Worley <address@hidden> wrote:

> From: Quentin Hartman <address@hidden>

> When running on a ceph-backed volume, I get closer to 15MB/s using the same
> tests, and have as much as 50% iowait. Typical operations that take seconds
> on bare metal take tens of seconds, or minutes in a VM. This problem
> actually drove me to look at things with strace, and I'm finding streams of
> FSYNC and PSELECT6 timeouts while the processes are running. More direct
> tests of ceph performance are able to saturate the nic, pushing about
> 90MB/s. I have ganglia installed on the host machines, and when I am
> running tests from within a vm ,the network throughput seems to be getting
> artificially capped. Rather than the more "spiky" graph produced by the
> direct ceph tests, I get a perfectly flat horizontal line at 10 or 20MB/s.

Is there some sort of virtio for networking?  It sounds like the
guest's driver thinks you've got a low-speed networking card.  I don't
see any options in the qemu command to control your networking
configuration.

Dale