| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-06-15 | |||
| 14:57:35 | mriedem | giblet: thanks | |
| 14:58:20 | mriedem | superdan: this is the failure http://logs.openstack.org/92/574792/5/check/neutron-grenade/7ec989c/logs/screen-n-cpu.txt.gz?level=TRACE#_Jun_14_16_51_14_918632 | |
| 14:58:23 | superdan | oh, it never ran the new side | |
| 14:58:28 | mriedem | yeah | |
| 14:58:39 | superdan | ack | |
| 14:58:41 | superdan | zcorneli: ^ | |
| 14:58:48 | mriedem | so....queens / old side is not using the queens UCA... | |
| 14:58:49 | superdan | mriedem: so I thought you said we had new enough stuff? | |
| 14:58:55 | mriedem | we do, in rocky | |
| 14:58:57 | mriedem | so, | |
| 14:59:05 | mriedem | we can also do a dnm backport of the queens uca devstack patch | |
| 14:59:06 | mriedem | sec | |
| 14:59:10 | superdan | heh | |
| 15:00:12 | zcorneli | The check did its job? | |
| 15:00:19 | superdan | yup | |
| 15:00:59 | mriedem | done | |
| 15:01:40 | mriedem | zcorneli: can you roll up the DNM nova patch on top of yours that depends on https://review.openstack.org/#/c/574792/ ? | |
| 15:01:47 | mriedem | zcorneli: do those words mean anything? | |
| 15:02:31 | mriedem | just need a nova patch on top of yours that has a depends-on in the commit message to https://review.openstack.org/#/c/574792/ - see how the depends-on in that change itself work | |
| 15:03:24 | zcorneli | The words mean things, not sure yet what the patch would contain. | |
| 15:03:36 | mriedem | just change a comment in nova/compute/manager.py to trigger tests | |
| 15:03:51 | mriedem | it's just a patch to link other changes together in the CI system | |
| 15:03:59 | mriedem | since devstack itself doesn't run the job we need | |
| 15:05:43 | zcorneli | Ah, so no real content to the patch, just a token change to trigger tests, with a depends-on the Use Queens UCA patch here: https://review.openstack.org/#/c/575764/ ? | |
| 15:06:15 | zcorneli | With a depends-on my patch as well? | |
| 15:07:24 | zcorneli | (I think I had it, and now I'm pretty sure I've confused myself. Not enough caffeine yet.) | |
| 15:09:29 | mriedem | wrong devstack change | |
| 15:09:44 | mriedem | zcorneli: depend on this https://review.openstack.org/#/c/574792/ | |
| 15:09:48 | mriedem | which depends on that other one | |
| 15:10:09 | mriedem | the dnm on top of your change doesn't require a depends-on to your change since they are in the same series / topic branch | |
| 15:10:34 | zcorneli | Ah. Got it now. | |
| 15:10:44 | zcorneli | Will do. | |
| 15:16:39 | openstackgerrit | Zack Cornelius proposed openstack/nova master: DNM: Libvirt file backed memory https://review.openstack.org/575766 | |
| 15:16:48 | zcorneli | mriedem: ^ | |
| 15:19:22 | giblet | finucannot: I replied in https://review.openstack.org/#/c/564092 | |
| 15:19:38 | finucannot | giblet: You sure? :) | |
| 15:20:02 | giblet | finucannot: friday. forget to push the button. It is there now | |
| 15:20:33 | finucannot | giblet: Oh, yeah. I'd link to the Rocky version and assume it will be there | |
| 15:20:49 | finucannot | giblet: More importantly though, are you happy to just get it in as is and fix after? | |
| 15:21:35 | giblet | finucannot: I can fix it now to point to the soon to be rocky version. I guess when the release notes are published that link won't be dead any more | |
| 15:21:49 | mriedem | giblet: finucannot: link to https://docs.openstack.org/nova/latest/reference/notifications.html#versioned-notification-samples | |
| 15:22:11 | giblet | mriedem: I can do that but finucannot doesn't like it | |
| 15:22:24 | mriedem | why? | |
| 15:22:31 | finucannot | mriedem: It'll break at some point in the future | |
| 15:22:45 | finucannot | giblet: To be clear, I'd prefer not to but I'm not firmly against it :) | |
| 15:22:48 | mriedem | i link to latest in the docs all the time | |
| 15:23:01 | mriedem | if it breaks in the future it's because we failed to add a redirect | |
| 15:23:01 | finucannot | and that's why we can't have nice things | |
| 15:23:36 | finucannot | Sections of docs do get moved around too, though I guess that's not a concern here | |
| 15:23:42 | finucannot | Yeah, use latest so | |
| 15:23:52 | openstackgerrit | Eric Fried proposed openstack/os-vif master: doc: Fix arg specs and object types in docs https://review.openstack.org/575151 | |
| 15:24:15 | openstackgerrit | Balazs Gibizer proposed openstack/nova master: Add full traceback to ExceptionPayload in versioned notifications https://review.openstack.org/564092 | |
| 15:24:18 | giblet | mriedem, finucannot: ^^ done | |
| 15:24:22 | mriedem | ack | |
| 15:25:37 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Add policy rule to block image-backed servers with 0 root disk flavor https://review.openstack.org/561284 | |
| 15:25:41 | mriedem | i'll trade you or anyone for ^ which is a security related issue (B1 classification) that's been sitting for a couple of months, and we know is impacting public clouds | |
| 15:26:02 | giblet | mriedem: looking | |
| 15:26:10 | finucannot | mriedem: I can take that if you can offer insight on https://review.openstack.org/#/c/575487/ | |
| 15:26:43 | finucannot | tl;dr: Without that patch, I find the RequestSpec.numa_networks field is unset when I get it at the scheduler (after being sent over RPC, I assume) | |
| 15:27:13 | finucannot | I assume something is off in the prior patch but I can't figure out what | |
| 15:27:40 | mriedem | is conductor converting the request spec to a primitive and back? | |
| 15:27:44 | mriedem | or legacy format and back? | |
| 15:27:48 | mriedem | there is some of that old cruft in conductor | |
| 15:28:25 | giblet | finucannot: could it be that you missing a similar piece than this https://review.openstack.org/#/c/567267/15/nova/objects/request_spec.py@486 ? | |
| 15:28:46 | finucannot | giblet: Oh, that could be it | |
| 15:29:03 | finucannot | *after* reviewing mriedem's patch | |
| 15:29:22 | giblet | finucannot: when RequestSpec.create() or RequestSpec.save() is called the code rewrites the RequestSpec object based on the db data | |
| 15:29:39 | giblet | finucannot: and your field is missing from the db data obviously | |
| 15:29:58 | mriedem | yeah my guess is _from_db_object is setting it to None | |
| 15:31:04 | mriedem | yup, that's it | |
| 15:31:10 | mriedem | RequestSpec.save() calls _from_db_object which nulls it out | |
| 15:31:14 | mriedem | pretty sure that's it | |
| 15:31:36 | mriedem | so follow this pattern i guess https://review.openstack.org/#/c/567267/15/nova/objects/request_spec.py@486 | |
| 15:32:37 | giblet | mriedem, finucannot: and when both feautre merges we can refactor RequestSpec to have a list of non persisted field names that automatically behave this way | |
| 15:33:20 | giblet | so the third developer introducing a non persisted field won't lose hair on this | |
| 15:43:55 | finucannot | mriedem: Done and yup, that was it. Cheers :) | |
| 15:44:24 | finucannot | giblet: agreed. I'll add mine as a list to ease that | |
| 15:47:01 | giblet | mriedem: I'm done with https://review.openstack.org/#/c/561284 as well. Overall +2 but both of us found small nits. | |
| 15:47:52 | mriedem | giblet: finucannot: question in https://review.openstack.org/#/c/564092/ | |
| 15:50:24 | mriedem | finucannot: giblet: thanks will respin that policy patch quick | |
| 15:52:22 | openstackgerrit | Merged openstack/nova master: Remove max_size parameter from fake_libvirt_utils.fetch_*image methods https://review.openstack.org/574911 | |
| 15:52:40 | giblet | mriedem: thanks for the review on the traceback in notification patch. | |
| 15:53:09 | giblet | mriedem: you have good points | |
| 15:53:13 | mriedem | np, Kevin_Zheng will string me up internally if i don't anyway :) | |
| 15:53:49 | giblet | mriedem: I don't have the time any more today to respin. Is it OK to let Kevin_Zheng respin it on Monday? | |
| 15:54:02 | mriedem | sure | |
| 15:56:38 | mriedem | finucannot: i'm not sure that .. warning:: works in the rendered policy docs | |
| 15:56:41 | mriedem | they aren't rst | |
| 15:58:00 | mriedem | giblet: are you ok with finucannot +Wing my change as a proxy to your +2 on that policy patch if i've addressed your comments? | |
| 15:58:02 | finucannot | mriedem: Oh, possible not. I guess I need to implement the same thing for that as I did for oslo.config so https://github.com/openstack/oslo.config/commit/03ff144f0 | |
| 15:58:08 | giblet | mriedem: sure | |
| 15:58:13 | mriedem | cool thanks | |
| 15:58:15 | giblet | mriedem: does are just nits | |
| 15:58:25 | finucannot | mriedem: ignore that comment, in that case | |
| 15:59:07 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Add policy rule to block image-backed servers with 0 root disk flavor https://review.openstack.org/561284 | |
| 15:59:08 | mriedem | done ^ | |
| 16:02:48 | melwitt | giblet: is https://blueprints.launchpad.net/nova/+spec/bandwidth-resource-provider stuck behind nested resource providers or no? (working on this etherpad https://etherpad.openstack.org/p/nova-rocky-blueprint-status) | |
| 16:03:17 | melwitt | kashyap: is https://blueprints.launchpad.net/nova/+spec/libvirt-cpu-model-extra-flags complete? it seems like all related changes have landed | |
| 16:03:37 | melwitt | kashyap: nevermind sorry, missed your reply in the logs | |
| 16:14:16 | melwitt | mriedem: noticed this bp is approved https://blueprints.launchpad.net/nova/+spec/servers-ips-non-unique-network-names but the spec isn't approved https://review.openstack.org/558125 | |