On Mon, 11 Aug 2014, Yuri Weinstein wrote: > I thought we could do the same in run-time for vps'es only. > > Sage? That's what I'm suggesting here.. move the change in that pull request into vps.yaml, and put vps.yaml inside ceph-qa-suite.git (maybe machine_types/vps.yaml?) so that it doesn't live in ~teuthworker? Need to confirm that that value can be set with an overrides, though ... sage > > On Mon, Aug 11, 2014 at 11:47 AM, Loic Dachary <loic@xxxxxxxxxxx> wrote: > > > > > > On 11/08/2014 19:34, Yuri Weinstein wrote: > >> Here is what we have in vps.yaml now: > >> > >> overrides: > >> ceph: > >> conf: > >> global: > >> osd heartbeat grace: 40 > >> > >> What do we want to add? > > > > I think the idle_timeout values at > > > > https://github.com/ceph/ceph-qa-suite/pull/79/files > > > > > >> > >> ~ > >> > >> On Mon, Aug 11, 2014 at 10:13 AM, Sage Weil <sweil@xxxxxxxxxx> wrote: > >>> On Mon, 11 Aug 2014, Yehuda Sadeh wrote: > >>>> Yeah, looking at these logs, it really seem that it's just that things > >>>> are going slow on these machines and it's hitting timeouts. The fix is > >>>> ok with me, although I'd rather have it adjusted per machine type > >>>> (somehow). > >>> > >>> There is a vps.yaml that bumps up another timeout, so we could put it > >>> there. Right now it lives on the teuthology machine > >>> (~teuthworker/vps.yaml I think?), but perhaps we should stick it in > >>> ceph-qa-suite.git somewhere ... > >>> > >>> sage > >>> > >>>> > >>>> Yehuda > >>>> > >>>> On Mon, Aug 11, 2014 at 9:21 AM, Loic Dachary <loic@xxxxxxxxxxx> wrote: > >>>>> Hi Yehuda, > >>>>> > >>>>> It looks like increasing the rgw idle timeout makes the problem go away ( https://github.com/ceph/ceph-qa-suite/pull/79 and http://tracker.ceph.com/issues/8988 ). It previously was 300 sec which looks like a large value already. Does this fix / workaround make sense to you ? > >>>>> > >>>>> Cheers > >>>>> > >>>>> On 10/08/2014 10:46, Loic Dachary wrote: > >>>>>> Hi Yehuda, > >>>>>> > >>>>>> In the past few months the swift tests failed randomly and I was unfortunately unable to figure out why. Here are a few examples: > >>>>>> > >>>>>> http://pulpito.ceph.com/loic-2014-08-08_12:17:30-upgrade:firefly-x:stress-split-wip-9025-chunk-remapping-testing-basic-vps/406944 > >>>>>> http://pulpito.ceph.com/loic-2014-08-08_12:17:30-upgrade:firefly-x:stress-split-wip-9025-chunk-remapping-testing-basic-vps/406941 > >>>>>> http://pulpito.ceph.com/loic-2014-08-08_12:17:30-upgrade:firefly-x:stress-split-wip-9025-chunk-remapping-testing-basic-vps/406946 > >>>>>> http://pulpito.ceph.com/loic-2014-08-08_12:17:30-upgrade:firefly-x:stress-split-wip-9025-chunk-remapping-testing-basic-vps/406947 > >>>>>> > >>>>>> and it has happened on every upgrade test run since I can remember. I fail to see a pattern and cannot figure out what the real problem is. It would be really great if you could take a look. Even a hunch or a tip would be greatly appreciated :-) > >>>>>> > >>>>>> You can find more context in > >>>>>> > >>>>>> http://tracker.ceph.com/issues/8988 > >>>>>> http://tracker.ceph.com/issues/8016 > >>>>>> http://tracker.ceph.com/issues/7799 > >>>>>> > >>>>>> and discussions at > >>>>>> > >>>>>> http://www.spinics.net/lists/ceph-devel/msg19933.html > >>>>>> > >>>>>> Cheers > >>>>>> > >>>>> > >>>>> -- > >>>>> Lo?c Dachary, Artisan Logiciel Libre > >>>>> > >>>> -- > >>>> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in > >>>> the body of a message to majordomo@xxxxxxxxxxxxxxx > >>>> More majordomo info at http://vger.kernel.org/majordomo-info.html > >>>> > >>>> > >>> -- > >>> To unsubscribe from this list: send the line "unsubscribe ceph-devel" in > >>> the body of a message to majordomo@xxxxxxxxxxxxxxx > >>> More majordomo info at http://vger.kernel.org/majordomo-info.html > > > > -- > > Lo?c Dachary, Artisan Logiciel Libre > > > -- > To unsubscribe from this list: send the line "unsubscribe ceph-devel" in > the body of a message to majordomo@xxxxxxxxxxxxxxx > More majordomo info at http://vger.kernel.org/majordomo-info.html > > -- To unsubscribe from this list: send the line "unsubscribe ceph-devel" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
