| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2019-11-19 | |||
| 16:21:24 | artom | Mostly the unit tests | |
| 16:24:01 | sean-k-mooney | i actully intionally did not add them but yes im going to add some for the neew fucntion | |
| 16:25:00 | sean-k-mooney | mainly because functional tests and exsiting test provide equal or better coverage in general but as noted ill add the test for the function you highlighted | |
| 16:29:31 | openstackgerrit | Thierry Carrez proposed openstack/nova master: Start README.rst with a better title https://review.opendev.org/695025 | |
| 16:32:35 | mgoddard | mriedem: hi. The more I pick at this thing (rebalance) the more it comes apart :) | |
| 16:32:43 | openstackgerrit | Merged openstack/nova master: (Temporarily) readd bare support for py27 https://review.opendev.org/695007 | |
| 16:33:11 | gmann | everyone, sent the ML for making devstack py3 default and other suggestion we agreed today- http://lists.openstack.org/pipermail/openstack-discuss/2019-November/010938.html | |
| 16:33:35 | mgoddard | mriedem: I have a few related but separable changes, and a functional tests. Does nova tend to prefer separate commits where possible, or shall I bung it all in one? | |
| 16:34:29 | mriedem | mgoddard: i can't speak for all of nova but i prefer separate commits if possible - easier to review, easier to revert if necessary | |
| 16:34:49 | mgoddard | mriedem: sure. | |
| 16:35:01 | mriedem | mgoddard: and i'm not surprised rebalance has issues since it's never had functional testing | |
| 16:35:16 | kashyap | mgoddard: On separate commits, this is a good reminder: https://wiki.openstack.org/wiki/GitCommitMessages#Structural_split_of_changes | |
| 16:35:24 | mriedem | the only way i found the race bug in the first place was digging into ironic multinode grenade logs looking for some other issue | |
| 16:35:29 | kashyap | (Bearing in mind things like bisectability, etc) | |
| 16:35:41 | mgoddard | mriedem: separate commits should be quite neat as each will fix a different part of the functional test | |
| 16:36:20 | bauzas | gibi: mriedem: do you think it would be worth inspecting the args when calling PlacementFixture ? | |
| 16:36:44 | bauzas | (related to the microversion/version issue) | |
| 16:37:15 | mgoddard | thanks kashyap. I generally prefer them, although it's not universal | |
| 16:39:32 | efried | bauzas: can you point me to the change in question please? | |
| 16:39:44 | efried | (It's somewhere on my list, been meaning to weigh in, but don't want to dig it up rn) | |
| 16:40:39 | efried | if this is a functional test, the PlacementFixture should be doing the right thing with its kwargs. And there should usually not be a reason to (re-)mock a method on the fixture. | |
| 16:40:44 | bauzas | efried: https://review.opendev.org/#/c/670112/8/nova/cmd/manage.py@2867 | |
| 16:41:10 | bauzas | context is that I used a wrong argument : 'microversion' instead of 'version' | |
| 16:41:39 | bauzas | so I could rather use utils.getargspec() | |
| 16:41:49 | efried | I understand, gimme sec. | |
| 16:41:53 | bauzas | I mean, in the PlacementFixture | |
| 16:42:45 | bauzas | efried: ie. something around there https://github.com/openstack/nova/blob/master/nova/tests/functional/fixtures.py#L103 | |
| 16:43:43 | efried | oh, crap, yeah, I remember trying to unwind this _fake_* gorp a while back and back-burnering it because it turned out to be hard. | |
| 16:44:37 | efried | yeah, this is (still) awful to have in the way of the actual placement fixture | |
| 16:44:52 | efried | it will mask all kinds of problems, like the one you're running into. | |
| 16:45:58 | efried | egads, we should also be making self._client a sdk proxy now as well. | |
| 16:46:07 | bauzas | you know what ? I'll provide a separate change | |
| 16:46:11 | bauzas | efried: ^ | |
| 16:46:15 | efried | not that it's hugely different | |
| 16:46:28 | efried | bauzas: Totally agree a separate change is in order. | |
| 16:46:34 | bauzas | k | |
| 16:47:08 | efried | If you can't completely kill the _fake_* methods, they should at least be avoiding massaging the args as much as possible. | |
| 16:47:21 | gmann | stephenfin: can you rebase this on top of slaweq patch to dropneutron-grenade from devstack (commented on review) - https://review.opendev.org/#/c/649097/10 | |
| 16:48:30 | stephenfin | ack | |
| 16:51:20 | openstackgerrit | Merged openstack/nova stable/pike: Revert "openstack server create" to "nova boot" in nova docs https://review.opendev.org/693236 | |
| 16:52:23 | stephenfin | gmann: done | |
| 16:55:10 | gmann | stephenfin: thanks, only thing we need to care is neutron-grenade job which will not work on devstack-py3-default. slaweq has patches up to migrate its usage to grenade-py3 but not all are merged. so i am making neutron-greande explicitly say py2 | |
| 16:56:03 | stephenfin | gmann: Makes sense. Do we need https://review.opendev.org/#/c/649097/ to depend on that so? | |
| 16:56:42 | gmann | stephenfin: yeah, i am pushing the neutron patch (neutron-grenade is in neutron repo) and ping you | |
| 16:58:11 | gibi | bauzas: tested your placement audit patch. The basic functionality seems to work with even bandwidth involved. But I think --delete is broken somehow. (See my comment in the review) | |
| 16:58:23 | bauzas | gibi: k, will look | |
| 16:58:37 | bauzas | gibi: and thanks :) | |
| 16:59:02 | gibi | np | |
| 16:59:34 | bauzas | gibi: I can't see your comment on delete ? | |
| 16:59:45 | bauzas | anyway, let's discuss this tomorrow | |
| 16:59:49 | bauzas | thanks gibi a lot | |
| 17:00:20 | bauzas | ah, nevermind | |
| 17:00:23 | bauzas | Ctrl-R | |
| 17:04:56 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Remove 'os-consoles' API https://review.opendev.org/687907 | |
| 17:04:56 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Remove 'nova-console' service, 'console' RPC API https://review.opendev.org/687908 | |
| 17:04:57 | openstackgerrit | Stephen Finucane proposed openstack/nova master: Remove 'nova-xvpvncproxy' https://review.opendev.org/687909 | |
| 17:59:41 | mriedem | dansmith: you want to move this bottom cross-cell patch along? https://review.opendev.org/#/c/642591/ - you should already be familiar with this one | |
| 18:10:57 | dansmith | mriedem: yeah, sorry I need to re-find my way on those | |
| 18:13:50 | mriedem | bauzas: easy bug fix https://review.opendev.org/#/c/694756/ | |
| 18:15:17 | openstackgerrit | Kashyap Chamarthy proposed openstack/nova master: libvirt: Bump MIN_{LIBVIRT,QEMU}_VERSION for "Ussuri" https://review.opendev.org/695056 | |
| 18:58:22 | openstackgerrit | Boris Bobrov proposed openstack/nova master: Also enable iommu for virtio controllers and video in libvirt https://review.opendev.org/684825 | |
| 19:19:49 | openstackgerrit | Merged openstack/nova master: FUP to Ie1a0cbd82a617dbcc15729647218ac3e9cd0e5a9 https://review.opendev.org/694787 | |
| 19:32:00 | efried | mriedem: Do you remember seeing/fixing this recently? https://zuul.opendev.org/t/openstack/build/ac433b3305ba454f845990f98a3bf481/log/logs/grenade.sh.txt.gz#38749 | |
| 19:33:30 | efried | I don't see it in e-r | |
| 19:45:44 | mriedem | efried: weren't you hitting something like that with trying to get osc to gate on py3? | |
| 19:46:05 | efried | this is what I don't remember. | |
| 19:46:21 | efried | fwiw this is trying to see whether barbican will gate on py3 | |
| 19:47:56 | mriedem | so it's probably installing py27 on the old side and then trying to upgrade to py3 on the new side and packages are getting mixed up | |
| 19:48:06 | mriedem | see https://review.opendev.org/#/c/694237/ | |
| 19:48:19 | mriedem | and https://github.com/openstack/grenade/blob/master/playbooks/legacy/grenade-py3/run.yaml | |
| 19:49:23 | mriedem | do the same in https://github.com/openstack/barbican/blob/master/playbooks/legacy/grenade-devstack-barbican/run.yaml | |
| 19:49:30 | mriedem | to make your barbican py3 dreams come true | |
| 19:51:55 | efried | mriedem: "the same" meaning add | |
| 19:51:55 | efried | export DEVSTACK_GATE_USE_PYTHON3=True | |
| 19:51:55 | efried | ? | |
| 19:52:34 | mriedem | yeah | |
| 19:52:44 | efried | ack, trying, thanks. | |
| 19:59:20 | dansmith | let it be known gmann_afk is a gentleman and his word is his bond | |
| 20:21:28 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Allow disabling cold migration to the same host https://review.opendev.org/676022 | |
| 20:26:45 | dustinc | dansmith: If you have a minute, would you mind taking a peek at this real quick and offering your opinion? https://review.opendev.org/#/c/693414/ | |
| 20:28:19 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Allow disabling cold migration to the same host https://review.opendev.org/676022 | |
| 20:30:06 | openstackgerrit | Eric Fried proposed openstack/nova-specs master: Spec: Ussuri: Encrypted Emulated Virtual TPM https://review.opendev.org/686804 | |
| 20:30:44 | efried | dansmith: Updated per your comments, but some of them left more questions than answers. I could use some help, please, when you get a chance. | |
| 20:31:17 | efried | jroll: ^ might come down to "are we dead in the water if we simply can't support shelve-offload (and possibly others)?" | |
| 20:31:44 | openstackgerrit | Artom Lifshitz proposed openstack/nova master: Functional tests for NUMA live migration https://review.opendev.org/672595 | |
| 20:31:46 | openstackgerrit | Artom Lifshitz proposed openstack/nova master: NUMA LM: Add func test for bug 1845146 https://review.opendev.org/687404 | |
| 20:31:46 | openstack | bug 1845146 in OpenStack Compute (nova) train "NUMA aware live migration failed when vCPU pin set" [High,Fix committed] https://launchpad.net/bugs/1845146 - Assigned to Dan Smith (danms) | |
| 20:34:45 | dansmith | dustinc: my opinion on what? whether it's fundamentally okay? | |
| 20:36:57 | dustinc | dansmith: Yes, efried suggested an alternative and I wanted help figuring out what makes more sense | |
| 20:39:17 | efried | dansmith: Trying to address what happens when the same node appears more than once in provider configs. Due to e.g. ironic rebalance we can't just fail, because the "conflict" might not appear until we're already running. So we have to set some kind of precedence rules for ``$COMPUTE_NODE`` vs explicit uuid and name | |
| 20:40:21 | dansmith | ohh, I clicked on the wrong link | |
| 20:40:39 | efried | In practice, I think this could only possibly be an issue at runtime in an ironic deployment when a rebalance happens, I think. At which point a new node will "appear" and we have to decide whether it has a provider config. | |
| 20:40:49 | dansmith | I have to run an errand, so I should probably not try to squeeze in a quick read of that before I go. | |
| 20:41:19 | efried | If we commit to this only being an issue in that ^ scenario, then I *think* the fact that ironic node RP name == uuid (doesn't it?) would let us hardcode an unambiguous solution pretty easily. | |
| 20:41:44 | efried | I need to food and then school run, will check back on this later. | |
| 20:41:58 | dansmith | I would punt to mriedem on that as I think he has the most ironic RP stuff in his head | |
| 20:42:03 | dansmith | but yeah, will leave this open for later | |
| 20:45:55 | openstackgerrit | Matt Riedemann proposed openstack/os-traits master: Add COMPUTE_SAME_HOST_COLD_MIGRATE trait https://review.opendev.org/666604 | |
| 20:47:36 | openstackgerrit | Merged openstack/nova master: Fix driver tests on Windows https://review.opendev.org/518525 | |
| 20:50:43 | mriedem | "ironic node RP name == uuid (doesn't it?)" yes | |