On 04/16/2013 02:08 AM, Wolfgang Hennerbichler wrote:
On 03/29/2013 09:46 PM, Josh Durgin wrote:
The issue was that the qemu rbd driver was blocking the main qemu
thread when flush was called, since it was using a synchronous flush.
Fixing this involves patches to librbd to add an asynchronous flush,
and a patch to qemu to use it.
The ceph patches are in the next and master branches, and will be
backported to bobtail. The patch for qemu is at [1] if anyone wants to
try it out.
Josh
[1] http://lists.gnu.org/archive/html/qemu-devel/2013-03/msg05367.html
Hi,
I can confirm that it performs VERY well now under load, no more I/O
hangs. Important for me was to set the VM-Type to pc-1.3 in libvirt,
like this:
<type arch='x86_64' machine='pc-1.3'>hvm</type>.
with my pc-1.0 setting before it was VERY slow (not ceph, but the
virtual system, this was a qemu-issue).
Hmm, good to know. Do you know what the qemu issue was?
The patch you posted to the qemu-mailinglist has not been included into
their master branch yet as far as I've seen. Are they reliable in
integrating it into upstream? This patch is REALLY relevant, IMHO, we
should urge them to integrate it sooner than later.
Yes, and it's included in the latest block pull request
(http://lists.gnu.org/archive/html/qemu-devel/2013-04/msg02706.html).
This and a bunch of other fixes are backported on top of bobtail in
the wip-bobtail-rbd-backports-req-order branch. The backport isn't
fully tested yet, but it's there if anyone wants to try it out.
Josh
_______________________________________________
ceph-users mailing list
ceph-users@xxxxxxxxxxxxxx
http://lists.ceph.com/listinfo.cgi/ceph-users-ceph.com
