| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-08-01 | |||
| 11:43:56 | sdague | gibi: I can actually bump it if you like, you are working on more important patches | |
| 11:44:54 | jaypipes | cdent: not on this bug, no. but in debugging, I did notice that the aggregate refresh rate in the resource tracker/report client was super high. I know you'd submitted a bug on that (or related) some time ago. might be nice to look at reducing the number of those calls. | |
| 11:44:57 | openstackgerrit | Sean Dague proposed openstack/nova master: do not pass proxy env variables by tox https://review.openstack.org/487327 | |
| 11:45:08 | sdague | gibi: fixed + approved | |
| 11:46:13 | openstackgerrit | Merged openstack/python-novaclient master: Remove run_tests.sh https://review.openstack.org/488276 | |
| 11:46:46 | cdent | jaypipes: it’s on the cycle of the periodic job isn’t it? Or is it happening even more than that? | |
| 11:47:09 | jaypipes | cdent: it happens every time reportclient._ensure_resource_provider() runs | |
| 11:47:14 | jaypipes | cdent: so.. a lot. | |
| 11:47:17 | gibi | sdague: thanks a lot! | |
| 11:47:46 | cdent | jaypipes: ah, okay, yeah. these days we do that even more. will poke. | |
| 11:51:40 | openstackgerrit | Merged openstack/python-novaclient master: Update URLs in docs, comments and setup.cfg https://review.openstack.org/488271 | |
| 12:00:14 | openstackgerrit | Michael Still proposed openstack/nova master: Move execs of tee to privsep. https://review.openstack.org/489438 | |
| 12:00:14 | openstackgerrit | Michael Still proposed openstack/nova master: Read from console ptys using privsep. https://review.openstack.org/489486 | |
| 12:26:11 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Test resize with placement api https://review.openstack.org/487958 | |
| 12:27:26 | gibi | sdague: I've fixed your comments in https://review.openstack.org/487958 | |
| 12:32:49 | mriedem | gibi: want to skip the notification meeting today? | |
| 12:33:17 | gibi | mriedem: I can chair it but I don't see any pressing issue to discuss | |
| 12:33:57 | gibi | mriedem: therefore I suggested to skip | |
| 12:34:05 | mriedem | seconded | |
| 12:34:34 | gibi | mriedem: Should I send a second mail to the ML about the skip? | |
| 12:34:43 | openstackgerrit | OpenStack Proposal Bot proposed openstack/nova master: Imported Translations from Zanata https://review.openstack.org/477091 | |
| 12:35:21 | mriedem | gibi: up to you | |
| 12:35:54 | gibi | mriedem: then I won't. I don't think anybody besides you and me could be interested | |
| 12:42:10 | jaypipes | gibi: gah. I was just in the process of reworking that resize tests... | |
| 12:45:42 | sdague | mriedem: ... http://logs.openstack.org/81/488381/3/experimental/gate-tempest-dsvm-cells-ubuntu-xenial/b5f4fa1/logs/devstacklog.txt.gz#_2017-08-01_10_56_05_959 ... oh cells munging hostnames | |
| 12:46:10 | sdague | cells v1 that is | |
| 12:47:18 | sdague | we should rename cells v1 to apoptosis | |
| 12:47:52 | mriedem | it's not deprecated yet | |
| 12:47:59 | mriedem | cells v1 is alive and well | |
| 12:48:06 | sdague | alive... sure | |
| 12:48:22 | sdague | mriedem: we really haven't deprecated it yet? | |
| 12:48:51 | sdague | because it's definitely in freeze for everything bug critical fixes, it seems like deprecation should be a thing | |
| 12:48:57 | sdague | even if we don't have a remove date | |
| 12:51:17 | gibi | jaypipes: sorry for that. I read back and only see an opportunity to refactor but not that somebody already started that | |
| 12:52:17 | mriedem | i'm +2 on gibi's test change | |
| 12:52:53 | mriedem | sdague: yes i've been thinking about deprecating cells v1 before we release pike, in case we want to drop it in queens | |
| 12:53:04 | mriedem | b/c we can't drop nova-network until we drop cells v1 and i don't want to lose another cycle on that | |
| 12:54:31 | jaypipes | mriedem: hold up. I'm fixing up a couple more things. | |
| 12:57:43 | openstackgerrit | Merged openstack/nova master: provide interface-scoped nameserver information https://review.openstack.org/467699 | |
| 13:02:32 | bauzas | jaypipes: gibi: I'm confused by the current situation for reviewing, how can I help by looking at first? | |
| 13:03:10 | jaypipes | bauzas: please give me 15 mins to push up another patch on top of gibi's latest test resize patch | |
| 13:03:18 | openstackgerrit | Rodolfo Alonso Hernandez proposed openstack/os-vif master: Read datapath_type from VIF object https://review.openstack.org/474914 | |
| 13:04:16 | bauzas | jaypipes: okay | |
| 13:04:32 | bauzas | jaypipes: I missed your convo yesterday with dan, hoping to understand more the issues | |
| 13:06:41 | mriedem | hooray for neighborhood construction taking out all the power | |
| 13:07:46 | dims | mriedem : yikes | |
| 13:09:11 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Don't warn on expected network-vif-unplugged events https://review.openstack.org/465794 | |
| 13:10:31 | jaypipes | bauzas: you didn't miss much other than me getting super raging. | |
| 13:13:52 | sdague | mriedem: https://review.openstack.org/#/c/487860/ that nova manage change, most of the urls you suggested in the test aren't valid urls | |
| 13:15:26 | openstackgerrit | Ildiko Vancsa proposed openstack/nova master: Implement new attach Cinder flow https://review.openstack.org/330285 | |
| 13:15:32 | mriedem | sdague: it's on the list | |
| 13:16:08 | sdague | mriedem: I also put a procedural hold on the privsep series once it started changing real things - https://review.openstack.org/#/c/459166/ | |
| 13:16:18 | sdague | I assume that's about right for this point in the release | |
| 13:16:24 | mriedem | yes | |
| 13:22:26 | mriedem | sdague: want to push this through? https://review.openstack.org/#/c/487932/ - i think what i'm hearing in there is we have to test it in prod | |
| 13:22:38 | mriedem | but if it works, we have other redirects to add, like for the api microversion history stuff | |
| 13:23:02 | mriedem | and those are just 2 that i know are broken, i don't know what else i don't know about | |
| 13:23:29 | openstackgerrit | Merged openstack/nova master: [placement] Add api-ref for RP traits https://review.openstack.org/474550 | |
| 13:23:56 | openstackgerrit | Merged openstack/nova master: doc: add FAQ entry for cells v1 config options https://review.openstack.org/487938 | |
| 13:24:13 | sdague | mriedem: looking | |
| 13:24:24 | openstackgerrit | Merged openstack/nova master: do not pass proxy env variables by tox https://review.openstack.org/487327 | |
| 13:24:44 | sdague | mriedem: yeh, +A | |
| 13:24:56 | sdague | we can test those redirects out in production and get them fixed there | |
| 13:25:05 | mriedem | stephenfin: are you going to update this? https://review.openstack.org/#/c/477497/ | |
| 13:25:16 | mriedem | stephenfin: i've identified some missing nova CLI guides from the admin guide that are missing in there | |
| 13:25:32 | openstackgerrit | Jay Pipes proposed openstack/nova master: Test resize with placement api https://review.openstack.org/487958 | |
| 13:25:32 | openstackgerrit | Jay Pipes proposed openstack/nova master: placement: remove existing allocs when set allocs https://review.openstack.org/489273 | |
| 13:25:33 | openstackgerrit | Jay Pipes proposed openstack/nova master: remove source provider allocs in confirm_resize() https://review.openstack.org/488510 | |
| 13:27:00 | jaypipes | mriedem, bauzas, gibi: I think you'll want to take a look at the changes I made to the test above. The tests now fail -- with no code changes -- indicating that the faulty assertions that supposedly were verifying the existing bad behaviour were incorrect. | |
| 13:28:57 | jaypipes | mriedem, bauzas, gibi: specifically, after the changes to the test I made, the following fails: | |
| 13:28:58 | mriedem | if you didn't change any function then why would they start failing? | |
| 13:29:00 | jaypipes | 1250 # NOTE(danms): This is bug 1707071 where we've lost the entire | |
| 13:29:00 | openstack | bug 1707071 in OpenStack Compute (nova) "Compute nodes will fight over allocations during migration" [Medium,In progress] https://launchpad.net/bugs/1707071 - Assigned to Jay Pipes (jaypipes) | |
| 13:29:00 | jaypipes | 1251 # allocation because one of the computes deleted it | |
| 13:29:00 | jaypipes | 1252 self.assertEqual(0, dest_usages['VCPU']) | |
| 13:29:14 | jaypipes | mriedem: because the test assertion was wrong. | |
| 13:29:22 | mriedem | then why wasn't it failing before? | |
| 13:30:03 | jaypipes | mriedem: I believe because the test was incorrectly passing the source rp UUID when it was supposed to pass the dest rp UUID | |
| 13:32:01 | mriedem | jaypipes: so you did change something functional | |
| 13:32:05 | mriedem | on which line? | |
| 13:32:47 | gibi | jaypipes: could you point out the faulty line in the test in the previous patchset, in 1252 seems worked the same way before than now | |
| 13:33:31 | openstackgerrit | Merged openstack/nova master: add a redirect for the old cells landing page https://review.openstack.org/487932 | |
| 13:33:53 | openstackgerrit | OpenStack Proposal Bot proposed openstack/nova master: Updated from global requirements https://review.openstack.org/489604 | |
| 13:37:39 | jaypipes | gibi: your guess is as good as mine. :( | |
| 13:38:02 | jaypipes | mriedem: I did not make any functional changes... :( | |
| 13:38:52 | cdent | jaypipes, gibi : you two didn’t collide on your rebasing/patching of the test did you? | |
| 13:38:56 | jaypipes | gibi, mriedem: the only thing I changes was not using this: | |
| 13:38:56 | jaypipes | server['OS-EXT-SRV-ATTR:host'] | |
| 13:38:57 | mriedem | jaypipes: ok, i don't know what you're saying then. you didn't make any changes, but things are failing, but they weren't before | |
| 13:39:13 | jaypipes | cdent: no, I rebased on a clean version from gibi | |
| 13:40:24 | jaypipes | mriedem, gibi: I believe the problem is the original test was relying on server['OS-EXT-SRV-ATTR:host'] to get the server's host and determine what was the "source" or "target". That's incorrect though because that value changes over the course of the migration | |
| 13:40:57 | jaypipes | so when I remove that and just use constants for the source and destination hostname (and rp UUIDs), the tests fails. | |
| 13:41:22 | mriedem | sdague: did you see my comments in here about versioning https://review.openstack.org/#/c/467699/ ? | |
| 13:41:22 | jaypipes | this is also why I think dansmith was seeing non-deterministic issues with ordering. | |
| 13:41:42 | jaypipes | because the server['OS-EXT-SRV-ATTR:host'] would change at different times in the migration sequence. | |
| 13:42:14 | dansmith | jaypipes: we got the host once at the beginning though | |
| 13:43:35 | jaypipes | dansmith: no... it was grabbed again on previous line 1215 in gibi's new _resize_and_check_allocations() | |
| 13:45:18 | jaypipes | dansmith: well, it wasn't "grabbed again"... just read again from the server dict. but the server dict is passed to the _wait_for_status() thing. perhaps that dict is modified? | |
| 13:45:30 | dansmith | jaypipes: that function wasn't there when Ilast pushed | |
| 13:45:44 | jaypipes | dansmith: I know, it was gibi's overnight refactor. | |