Index ¦ Archives ¦ Atom

Placement Update 18-47

Posted on: Fri 23 November 2018

Category: openstack – Tags: placement

It's been a while since the last placement update. Summit happened. Seemed pretty okay, except for the food. People have things they'd like to do with placement.

Most Important

We're starting to approach the point where we're thinking about the possibility of maybe turning off placement-in-nova. We're not there yet, and as is always the case with these kinds of things, it's the details at the end that present the challenges. As such there are a mass of changes spread around nova, placement, devstack, grenade, puppet and openstack-ansible related to making things go. More details on those below, but what we need is the same as ever: reviews. Don't be shy. If you're not a core or not familiar with placement, reviews are still very helpful. A lot of the patches take the form of "this might be the right way to do this".

What's Changed

There is now a placement-manage command which can do database migrations, driven by alembic. This means that the devstack patch which uses the extracted placement can merge soon. Several other testing related (turning on tempest and grenade for placement) changes depend-on that.

Matt did a placement-status command which has a no-op we-are-here upgrade check. We've already met the python3 goals (I think?), so I reckon placement is good to go on community-wide goals. Woot.

The PlacementFixture that placement provides for other projects to do functional tests with it has merged. There's a patch for nova using it.

The spec for counting quota usage in placement has been revived after learning at summit that a proposed workaround that didn't use placement wasn't really all that good for people using cells v2.

Bugs

Specs

Summit and U.S. Thanksgiving has disrupted progress on some of these, but there are still plenty of specs awaiting their future.

Many of these have unaddressed negative review comments.

Main Themes

Making Nested Useful

Progress is being made on gpu-reshaping for libvirt and xen:

Making use of nested is bandwidth-resource-provider:

Somewhat related to nested are a stack of changes to how often the ProviderTree in the resource tracker is checked against placement, and a variety of other "let's make this more right" changes in the same neighborhood:

Extraction

(There's an etherpad which tracks some of the work related to extraction. Please refer to that for additional information.)

TripleO and OpenStack-Ansible are both working on tooling to install and/or upgrade to extracted placement:

libvirt support for GPU reshaping:

Grenade and tempest testing for extracted placement:

A replacement for placeload performance testing that was in the nova-next job: https://review.openstack.org/#/c/619248/. This might be of interest to people trying to do testing of live services without devstack. It starts with a basic node, turns on mysql, runs placement with uwsgi, and does the placeload testing. Note that this found a pretty strange bug in _ensure_aggregates.

Documentation tuneups:

We've been putting off making a decision about os-resource-classes. Anyone have strong opinions?

Other

Besides the 20 or so open changes in placement itself, and those mentioned above, here are some other changes that may be of interest.

End

Lot going on. Thanks to everyone for their contributions.

Posted on: Fri 23 November 2018

Category: openstack – Tags: placement

© Chris Dent. Built using Pelican. Theme by Giulio Fidente on github.