On Fri, May 6, 2016 at 12:50 AM, Matthew Booth <[email protected]> wrote:
> I mentioned in the meeting last Tuesday that there are now 2 of us working > on the persistent storage metadata patches: myself and Diana Clarke. I've > also been talking to Paul Carlton today trying to work out how he can get > moving with the subsequent libvirt storage pools work without a huge amount > of conflict. We're going to work that out, but something which would help > us enormously while we work together is to knock off some semi-dependent > patches at the beginning of the series. > > Specifically there are 5 patches which precede the series. None of these 5 > patches are really part of the series, but the series depends on all 5 to a > greater or lesser degree. In order, they are: > > Only attempt to inject files if the injection disk exists > https://review.openstack.org/#/c/250872/ > > Remove fake_imagebackend.Raw and cleanup dependent tests > https://review.openstack.org/#/c/267661/ > > Rename Image.check_image_exists to Image.exists() > https://review.openstack.org/#/c/270998/ > > Rename Raw backend to NoBacking > https://review.openstack.org/#/c/279626/ > > Remove deprecated option libvirt.remove_unused_kernels > https://review.openstack.org/#/c/265886/ > > These have all been reviewed previously and attracted +1s. I just rebased > them, so they may have been lost, but I'm pretty sure they're ready for > prime time. These patches aren't really what the series is about, but they > are currently the principal source of merge conflicts. If we could get > these out of the way it would make it much easier for the 3 of us working > on the substantive series. Any chance of some core reviewer attention to > get these moving? > > As for the status of the rest of the series, there are 2 more which I > don't expect to change: > > Add a lock() context manager to image backend > https://review.openstack.org/#/c/279625/ > > Introduce ImageCacheLocalPool > https://review.openstack.org/#/c/279669/ > > Please review these too. However, these patches aren't ready to be merged > because they don't add users of the interfaces they introduce. > > Everything after that is definitely changing. Diana and I are currently > working on cranking through these backend by backend. I'll provide a weekly > progress update in the live migration meeting. > > TL;DR Core reviewers: please review the first 5 patches listed above. > There will be cake. > I'm very open to bribes. How will this cake be delivered? Michael -- Rackspace Australia
__________________________________________________________________________ OpenStack Development Mailing List (not for usage questions) Unsubscribe: [email protected]?subject:unsubscribe http://lists.openstack.org/cgi-bin/mailman/listinfo/openstack-dev
