| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-08-11 | |||
| 09:27:21 | openstackgerrit | yfzhao proposed openstack/python-novaclient master: Fix wrong links https://review.openstack.org/492921 | |
| 09:34:41 | gibi | good morning | |
| 09:39:41 | bauzas | gibi: good lunch :p | |
| 09:41:47 | gibi | bauzas: the day started earlier a bit but I was occupied with internal meetings | |
| 09:42:03 | gibi | bauzas: but yes, a lunch is about to happen too | |
| 09:42:42 | gibi | bauzas: did Matt cut rc1 last night? | |
| 09:43:12 | bauzas | gibi: no | |
| 09:43:23 | bauzas | gibi: https://review.openstack.org/#/c/492788/2 | |
| 09:43:39 | bauzas | he was waiting for the merge IIRC | |
| 09:43:53 | bauzas | gibi: now, the main problem is about the Ironic job | |
| 09:43:57 | bauzas | dtantsur: around ? | |
| 09:44:25 | bauzas | dtantsur: context is https://review.openstack.org/#/c/487954/14 | |
| 09:44:37 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: replace chance with filter scheduler in func tests https://review.openstack.org/491529 | |
| 09:44:52 | bauzas | dtantsur: that change is atm a RC1 blocker but I don't see the Ironic job providing the log saying that the flavor is modified | |
| 09:45:06 | gibi | bauzas: thanks for the info | |
| 09:45:33 | bauzas | dtantsur: if we cut RC1 without ^, that could possibly be an upgrade problem for Queens | |
| 09:45:55 | bauzas | so I would appreciate any ironic folks support now :) | |
| 09:46:06 | bauzas | TheJulia: around too ? | |
| 09:47:58 | smcginnis | bauzas: Do you need to hold that RC request? | |
| 09:49:34 | smcginnis | bauzas: Oh, nevermind. Now I see your comment there. | |
| 09:50:04 | bauzas | smcginnis: yeah, I'd like to discuss with matt about whether we cut RC1 without the change I provided in my comment | |
| 09:50:21 | bauzas | smcginnis: do you accept that ? is it possible to cut RC1 only on Friday? | |
| 09:50:30 | bauzas | I mean, it's also a branch | |
| 09:50:41 | smcginnis | bauzas: I think we're probably fine waiting until Monday if we have to. | |
| 09:50:49 | smcginnis | bauzas: You're just waiting on one patch, right? | |
| 09:51:05 | bauzas | smcginnis: yup, I just rebased the releases tag with the merge sha1 | |
| 09:51:32 | bauzas | smcginnis: so if we just cut the tag without the above change, a +W would only be needed | |
| 09:51:51 | bauzas | the only blocker for us is https://review.openstack.org/#/c/487954/ | |
| 09:52:18 | bauzas | other bugs would be treated in other RCs | |
| 09:52:34 | bauzas | HTH | |
| 09:52:36 | smcginnis | bauzas: Cool, I think we're fine waiting a little then. | |
| 09:55:27 | dtantsur | morning bauzas | |
| 09:55:52 | dtantsur | I'll start looking in it in a few minutes | |
| 10:00:23 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: test server evacuation with placement https://review.openstack.org/492548 | |
| 10:03:15 | bauzas | dtantsur: FWIW, I'm forging a DNM patch for adding more voerbosity | |
| 10:04:29 | dtantsur | cool | |
| 10:16:55 | openstackgerrit | Sylvain Bauza proposed openstack/nova master: Make compute log less verbose with allocs autocorrection https://review.openstack.org/492571 | |
| 10:18:56 | dtantsur | bauzas: I don't really understand the CI failure there.. it seems unrelated to me | |
| 10:19:19 | bauzas | dtantsur: hence me logging more | |
| 10:19:27 | bauzas | dtantsur: I'm even not sure we run the thread | |
| 10:19:43 | bauzas | given it's in a conditional | |
| 10:20:55 | dtantsur | we should not, it only runs for nodes with resource_class != None | |
| 10:21:25 | dtantsur | bauzas: I can make https://review.openstack.org/#/c/491777/ depends-on this patch, to see how it behaves with resource_classes. wdyt? | |
| 10:22:14 | bauzas | dtantsur: sure | |
| 10:22:43 | bauzas | dtantsur: if that doesn't trample you waiting for devstack patch being merged | |
| 10:23:00 | bauzas | dtantsur: I mean, if you can wait for that devstack patch to be merged, that's fine to me | |
| 10:23:01 | dtantsur | we need to make sure it won't blow up after merging the both changes :) | |
| 10:23:06 | bauzas | yeah | |
| 10:23:32 | bauzas | anyway, just doing an urgent internal bug thingy and then I'm back to the DNM patch | |
| 10:23:43 | dtantsur | ack | |
| 10:27:44 | dtantsur | ok, both patches updated, waiting for the CI | |
| 10:41:06 | openstackgerrit | Dmitry Tantsur proposed openstack/nova master: Deprecate bare metal filters https://review.openstack.org/492563 | |
| 10:42:04 | dtantsur | bauzas: meanwhile, do you think we can also get ^^ in? | |
| 10:53:11 | bauzas | dtantsur: looks to me hard for RC1 | |
| 10:53:21 | bauzas | dtantsur: even if we haven't yet merged it | |
| 10:53:26 | bauzas | tagged it, sorry | |
| 10:53:50 | dtantsur | ok, that's fine. I just have an ironic docs patch depending on it, I may need to split it | |
| 10:53:53 | bauzas | dtantsur: we're already past the deadline but I leave matt make the hard call :) | |
| 10:54:29 | dtantsur | vdrok: first of all, please review https://review.openstack.org/#/c/487954/ | |
| 10:54:52 | vdrok | dtantsur: looking | |
| 10:56:09 | dtantsur | vdrok: it fails the ironic CI for some reason. I see network connection problems between various services, so it is not necessary related to the patch itself | |
| 10:56:36 | vdrok | dtantsur: yup, there is some socket error in the vbmc log as well | |
| 10:56:51 | dtantsur | I've rechecked it, let's see | |
| 10:57:16 | dtantsur | vdrok: our next step would be to make https://review.openstack.org/491777 and https://review.openstack.org/476968 pass the CI - reviews welcome there too | |
| 10:57:26 | bauzas | dtantsur: I'm more concerned by the fact I don't see the logs mentioning the flavor update rather than the Ironic job giving us -1 :) | |
| 10:57:57 | bauzas | dtantsur: in other words, I feel brave enough to +2 some ironic change if I'm sure the job issues are unrelated | |
| 10:58:17 | dtantsur | bauzas: why should we see any updates, given that the nodes don't have resource classes yet? | |
| 10:58:38 | bauzas | oh f**** | |
| 10:59:01 | bauzas | dtantsur: you killed me :p | |
| 10:59:22 | dtantsur | bear metal powerzzz! | |
| 10:59:25 | bauzas | dtantsur: those ironic nodes aren't having resource classes | |
| 10:59:26 | bauzas | ? | |
| 10:59:31 | bauzas | yet, I mean ? | |
| 11:00:04 | dtantsur | bauzas: yep. your logging line should show up in https://review.openstack.org/491777 instead - hence I made it depends-on the nova patch | |
| 11:01:20 | bauzas | dtantsur: oh snap https://review.openstack.org/#/c/491777/9/devstack/lib/ironic@1821 right? | |
| 11:01:48 | bauzas | until that devstack change, the gate nodes aren't yet correctly having resource classes | |
| 11:01:55 | bauzas | I thought it was already the case | |
| 11:02:08 | bauzas | dtantsur: IMHO, we should invert the depends-on | |
| 11:02:10 | dtantsur | hah, sorry for not figuring out the confusion earlier | |
| 11:02:48 | bauzas | dtantsur: why would you make the devstack change dependent on the nova change ? | |
| 11:02:52 | dtantsur | bauzas: yeah, good call probably. wanna me drop the depends-on from my patch? | |
| 11:03:10 | bauzas | if the nova change uses what's provided by the devstack one ? | |
| 11:03:10 | dtantsur | I wanted one of them to depend on the other, I don't care which exactly :) | |
| 11:03:29 | bauzas | dtantsur: yeah, please remove the depends-on on the devstack one | |
| 11:03:51 | bauzas | dtantsur: and then I'll update https://review.openstack.org/#/c/487954/ to include devstack | |
| 11:04:12 | dtantsur | bauzas: done | |
| 11:04:20 | bauzas | dtantsur: I have to apologize, I wasn't having a full view of the situation | |
| 11:04:26 | vdrok | dtantsur: so, for that code to be triggered, we have to have an active instance booted with old flavor, and afterwards being updated with resource class right? | |
| 11:04:35 | bauzas | dtantsur: okay, I'm on https://review.openstack.org/#/c/487954/ | |
| 11:04:36 | vdrok | code in https://review.openstack.org/#/c/487954/14 I mean | |
| 11:04:36 | dtantsur | bauzas: no problem, thanks for helping us with this stuff anyway | |
| 11:05:08 | bauzas | vdrok: for the nova code to be triggered, you have to set resource classes for ironic nodes firsrt | |
| 11:05:29 | bauzas | vdrok: that would be done by devstack in the job we discuss | |
| 11:06:15 | dtantsur | vdrok: I think vdrok's point is that we still won't see the log message, because it needs the resource_class to not be present initially.. | |
| 11:06:30 | vdrok | dtantsur: bauzas exactly | |
| 11:06:50 | openstackgerrit | Sylvain Bauza proposed openstack/nova master: Handle addition of new nodes/instances in ironic flavor migration https://review.openstack.org/487954 | |
| 11:06:57 | dtantsur | so I wonder if the only option to test it is to actually get a devstack locally, and try it | |
| 11:07:02 | bauzas | dtantsur: done ^ | |
| 11:07:31 | bauzas | dtantsur: just to make it clear, I'm just updating it but just for testing purposes | |
| 11:07:52 | dtantsur | vdrok: in any case, could you please review https://review.openstack.org/#/c/491777/ ? this is something we must get in today to not block nova further | |