[openstack-dev] [Nova] RC1 state of play

Sylvain Bauza sbauza at redhat.com
Thu Apr 9 08:18:09 UTC 2015

Le 09/04/2015 08:01, Michael Still a écrit :
> There are a few bugs still outstanding for nova's RC1. Here's a quick
> summary. For each of these we need to either merge the fix, or bump
> the bug from being release blocking.
> -----
> https://bugs.launchpad.net/nova/+bug/1427351
> "cells: hypervisor API extension can't find compute_node services"
> This still has review https://review.openstack.org/#/c/160506/
> outstanding, but a related review has landed. Do we need to land the
> outstanding review as well?

We're in a good shape for merging the hypervisor-related issues that are 
fixed in https://review.openstack.org/#/c/160506/ so yes we need to land 
it too.

Unforunately, the cells job recently regressed due to some Tempest 
changes ([1] and others) related to networks, so there are a couple of 
last-minute patches that have been uploaded :
https://review.openstack.org/171865 for fixing the devstack side - 
creating a network in the child cell
https://review.openstack.org/171911 and 
https://review.openstack.org/171912 for fixing the Nova side because it 
has a miss (not using the UUID passed in the CLI if provided)

Both corresponding cell-related bugs are raised as Critical and marked 
for RC1 :

[1] https://github.com/openstack/tempest/commit/4bbc199

> -----
> https://bugs.launchpad.net/nova/+bug/1430239
> "Hyper-V: *DataRoot paths are not set for instances"
> This one has https://review.openstack.org/#/c/162999 proposed as a
> fix, which has one +2. Does anyone want to review a Hyper-V driver
> change?
> -----
> https://bugs.launchpad.net/nova/+bug/1431291
> "Scheduler Failures are no longer logged with enough detail for a site
> admin to do problem determination"
> Two reviews outstanding here --
> https://review.openstack.org/#/c/170421/ and its dependent (and WIP)
> https://review.openstack.org/#/c/170472/ -- these seem to be not
> really ready. What's the plan here?

Unscoped from RC1 milestone as it probably requires some big behaviour 
change for the scheduler RPC API (not raising a NoValidHost exception 
but rather return a list of filters failing)


> -----
> https://bugs.launchpad.net/nova/+bug/1313573
> "nova backup fails to backup an instance with attached volume
> (libvirt, LVM backed)"
> For this we've merged a change which raises an exception if you try to
> do this, so I think this is no longer release critical? It's still a
> valid bug though so this shouldn't be closed.
> -----
> https://bugs.launchpad.net/nova/+bug/1438238
> "Several concurent scheduling requests for CPU pinning may fail due to
> racy host_state handling"
> The fix is https://review.openstack.org/#/c/169245/, which needs more reviews.
> Michael

More information about the OpenStack-dev mailing list