| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-08-03 | |||
| 12:52:50 | edleafe | sambetts: looking at the bug report, it seems that the the flavor itself in wrong | |
| 12:53:13 | sambetts | edleafe: the bug report was for a different issue | |
| 12:53:14 | edleafe | sambetts: it zeroes out the "standard" resources, but doesn't include the ironic resource class | |
| 12:53:29 | sambetts | edleafe: the bug report is unrelated to using it with Ironic | |
| 12:53:42 | stephenfin | sdague: They're all done. The only one I'm not gone on is the last. I can fix the issues I have if you want https://review.openstack.org/#/c/490455/2 | |
| 12:53:43 | edleafe | sambetts: well, not really | |
| 12:54:00 | edleafe | as Ironic would be the only system using flavors like that | |
| 12:54:33 | sambetts | but in a non-ironic deployment there is nothing stopping me creating a flavor like that and nova doesn't error nicely | |
| 12:54:46 | sambetts | thats what the bug report is saying | |
| 12:55:02 | sambetts | thats why it doesn't mention ironic | |
| 12:55:35 | sambetts | you can't say its not a bug because someone would just not do that in a normal case | |
| 12:55:50 | edleafe | sambetts: Sure, but there are a zillion incorrect ways to do things. We can't protect against every wrong setup | |
| 12:56:51 | edleafe | sambetts: it is returning a 400, as it should | |
| 13:00:42 | sambetts | I just don't expect that nova should even make the request to the placement API, because the error message identifies a bad request from nova | |
| 13:07:39 | cdent | \o/ | |
| 13:08:35 | cdent | so no matt, jay or dan yet it seems, I guess I’ll see if can fix the test issues in 488510 | |
| 13:18:48 | edleafe | cdent: there are some ironic resource class issues if you're bored :) | |
| 13:19:23 | cdent | edleafe: heh, issues abound | |
| 13:19:33 | cdent | i’m looking at fixing these test issues | |
| 13:19:37 | cdent | they are mostly mechanical | |
| 13:19:42 | cdent | but plentiful | |
| 13:19:43 | edleafe | cool | |
| 13:29:12 | openstackgerrit | Stephen Finucane proposed openstack/nova master: reflow rpc doc to 80 columns https://review.openstack.org/490455 | |
| 13:30:34 | openstackgerrit | Matt Riedemann proposed openstack/nova master: hardware offload support for openvswitch https://review.openstack.org/398265 | |
| 13:37:12 | openstackgerrit | Merged openstack/nova master: Fix lists in process doc https://review.openstack.org/490431 | |
| 13:40:31 | dansmith | mriedem: are you working on that set? | |
| 13:40:33 | openstackgerrit | Merged openstack/nova master: Fix whitespace in rest_api_version_history https://review.openstack.org/490435 | |
| 13:40:43 | dansmith | mriedem: I just commented on both the bottom two | |
| 13:41:14 | openstackgerrit | Merged openstack/nova master: Fix list rendering in code-review.rst https://review.openstack.org/490437 | |
| 13:42:52 | openstackgerrit | Stephen Finucane proposed openstack/nova master: doc: Make use of definition lists, literals https://review.openstack.org/490481 | |
| 13:43:12 | cdent | dansmith, mriedem I was going to fix the failing unit tests on the top patch, but my lack of familiarity is meaning that teasing out the mocks is proving slow. If one of you two want to pick it up, that’s cool, otherwise I’ll carry on and learn something | |
| 13:43:34 | stephenfin | mriedem: Is it too late to ask for FFE for this? https://review.openstack.org/#/c/453808/3 | |
| 13:44:06 | dansmith | cdent: okay I'm surprised mriedem or jaypipes aren't already in the middle of that, so let's confirm before I go make another mess | |
| 13:44:06 | stephenfin | I can't work on moving nova-manage to cliff for another six months without it | |
| 13:46:08 | cdent | at least one of the failures is because we’ve now got code within methods making demands that unrelated “unit” tests are not satisfying | |
| 13:46:41 | cdent | (just the first one, the others make some sense) | |
| 13:51:25 | cdent | dansmith: are you jumping, or shall I carry on with what I’m able to get for now? | |
| 13:52:03 | dansmith | cdent: I'm handling other morning stuff waiting for mriedem to catch me up.. if you're fixing things, I'd just keep going with that | |
| 13:52:11 | dansmith | and/or push fixes you already have working | |
| 13:52:12 | cdent | roger that | |
| 13:52:20 | cdent | check with me before you dive in | |
| 13:52:25 | cdent | I’ll get what I can make sense of | |
| 13:52:50 | dansmith | ack | |
| 13:53:31 | jaypipes | I could use a catch-up too. | |
| 13:55:58 | openstackgerrit | Merged openstack/nova master: fix list rendering in rpc doc https://review.openstack.org/490442 | |
| 13:56:41 | openstackgerrit | Merged openstack/nova master: Fix list rendering in bdm doc. https://review.openstack.org/490446 | |
| 13:57:23 | openstackgerrit | Merged openstack/nova master: fix list rendering in aggregates https://review.openstack.org/490448 | |
| 13:58:12 | openstackgerrit | Merged openstack/nova master: fix list rendering in cells https://review.openstack.org/490449 | |
| 13:58:59 | openstackgerrit | Merged openstack/nova master: fix list rendering in policy-enforcement https://review.openstack.org/490451 | |
| 14:00:42 | dansmith | jaypipes: I had just said I was waiting to see if mriedem had anything queued locally for that set, I assume that means you do not? | |
| 14:02:17 | dansmith | jaypipes: per my comment a few minutes ago, if we don't address the flavor overrides in that un-doubling, we'll mess up accounting after a move with an overriding flavor | |
| 14:02:28 | dansmith | and edleafe has a flavor-to-resources method we can just use I think | |
| 14:10:42 | mriedem | dansmith: cdent: just got home, | |
| 14:10:48 | mriedem | i'm not currently working on anything | |
| 14:11:15 | dansmith | mriedem: okay | |
| 14:11:36 | dansmith | cdent: so I guess push your changes when you're ready and then I will see about fixing the resource override stuff | |
| 14:12:17 | openstackgerrit | Chris Dent proposed openstack/nova master: remove provider allocs in confirm/revert resize https://review.openstack.org/488510 | |
| 14:12:24 | cdent | good timing ^ | |
| 14:12:45 | cdent | that fixes the pep8 problems, and the calls where flavor was not included in a call to remove_provider_from … | |
| 14:13:12 | cdent | but it does not fix an assertion with AssertionError: Expected 'delete_allocation_for_instance' to be called once. Called 0 times. | |
| 14:13:23 | cdent | nor some mock problems in “unrelated” tests | |
| 14:14:35 | cdent | dansmith, mriedem, jaypipes, I release the lock | |
| 14:15:44 | dansmith | ah, mriedem's patch should already be kosher in this respect actually | |
| 14:15:53 | dansmith | I was thinking with my uncaffeinated brain | |
| 14:16:07 | dansmith | it's just the last one that needs it | |
| 14:16:38 | cdent | just to highlight it, gibi pointed out yet another edge case on ps16: https://review.openstack.org/#/c/490461 | |
| 14:17:32 | mriedem | dansmith: so no resource override in my change? | |
| 14:17:42 | mriedem | b/c if so, you'll have to deal with it b/c resource override is lost on me | |
| 14:17:44 | dansmith | mriedem: right yours should be good because the scheduler has already done that part | |
| 14:17:52 | mriedem | ok | |
| 14:22:50 | jaypipes | dansmith: sorry, having some issues today :( | |
| 14:23:26 | jaypipes | dansmith: last I checked in I was waiting on the bottom patch in that series to test positive and then I was going to appove it (it was the test resize same host patch) | |
| 14:24:03 | jaypipes | dansmith: I think my status currently is I need to pull whatever the latest that mriedem pushed last night and cherry-pick the changes I was working on last night onto that series. | |
| 14:24:19 | dansmith | okay that's what I was just asking, if you had pending changes | |
| 14:24:21 | jaypipes | dansmith: those changes include fixing up the "cheating" section | |
| 14:24:35 | dansmith | anyway, let me finish what I'm doing here, which is fixing the cheating section, but so it accounts for flavor overrides | |
| 14:24:37 | dansmith | almost done with that | |
| 14:24:44 | jaypipes | dansmith: oh, ok | |
| 14:24:47 | jaypipes | dansmith: np | |
| 14:25:03 | jaypipes | dansmith: and sorry, I'm having some medical issues today. trying to get in to see a doctor. | |
| 14:25:16 | dansmith | ack, np | |
| 14:25:21 | jaypipes | so it's not an ideal work day unfortunately :( | |
| 14:26:08 | jaypipes | if cdent would send me a box of Crunchie bars and a new hip and lower back, that would be groovy. | |
| 14:26:30 | jaypipes | thx cdent :) | |
| 14:26:52 | cdent | they are a part of my official travel to openstack events habit, so if you can survive until september: hooked up | |
| 14:27:08 | mriedem | *cortisone | |
| 14:27:09 | mriedem | damn | |
| 14:27:10 | cdent | I’m going home, brb | |
| 14:27:27 | openstackgerrit | Matthew Booth proposed openstack/nova master: imagebackend: cleanup constructor args to Rbd https://review.openstack.org/490499 | |
| 14:27:47 | jaypipes | mriedem: that cortisone come with morphine? | |
| 14:28:04 | mdbooth | ^^^ is the most trivial trivial patch I've ever submitted, btw | |
| 14:28:38 | mdbooth | Actually, I'm sure I've submitted a typo fix in a comment before | |
| 14:28:38 | openstack | Launchpad bug 1708424 in OpenStack Compute (nova) "When a flavor has resource extra_specs disabling all standard fields, nova tries to make a request to the placements API with no resources" [Undecided,New] | |
| 14:28:38 | sdague | https://bugs.launchpad.net/nova/+bug/1708424 ... that's a weird bug | |
| 14:29:01 | edleafe | sdague: IMO that one is operator error | |
| 14:29:08 | sdague | yeh | |
| 14:29:32 | edleafe | sdague: they go to the trouble of zeroing out all the standard resources, but never add the ironic resource class | |
| 14:30:05 | edleafe | sdague: so perhaps we could catch that in Nova, but if they really wanna shoot themselves in the foot... | |
| 14:30:11 | sdague | yeh | |
| 14:30:31 | sdague | so is that a hack to do something ironicy? | |