On 02/22/2018 02:06 PM, Sage Weil wrote: > On Thu, 22 Feb 2018, Sage Weil wrote: >> On Thu, 22 Feb 2018, Ken Dreyer wrote: >>> On Thu, Feb 22, 2018 at 2:32 PM, Alfredo Deza <adeza@xxxxxxxxxx> wrote: >>>> On Thu, Feb 22, 2018 at 4:26 PM, Ken Dreyer <kdreyer@xxxxxxxxxx> wrote: >>>>> Hi Alfredo and Kefu, >>>>> >>>>> I'm thinking about this in the context of the earlier ceph-devel >>>>> discussions about releases. >>>>> >>>>> Is there a particular difference between Jenkins' release builds and >>>>> Jenkins CI builds that caused us to miss this error until release day? >>>> >>>> There isn't a difference. The problem is that we don't build anything >>>> except CentOS7 and Xenial for CI >>> >>> I see what you mean now. Would it help if we built all the "release" >>> platforms weekly or something? >> >> And/or, is there a way to trigger this explicitly for a pr in jenkins >> (e.g., 'jenkins build all targets please')? >> >>> Not trying to sign you up for more work of course, just trying to >>> understand the scope of this. >> >> FWIW I don't think the ceph-deploy test would take much to fix up. My >> recollection is that one of the facets specifies the distro and it's just >> a matter of keeping the symlinks in ceph.git/qa/distros/supported/ in sync >> with the builds we want to test. > > Actually, I think the hard part before was make sure we had downburst > images for those distros. In the current/new world, I think it is now a > matter of ensuring that ovh has images for those distros? (I'm not quite > sure how the newer ovh/openstack teuthology integration works.) > > sage In this case, and in several, even a "build only" policy would save last-minute panics, right?... -- 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
