| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2018-03-16 | |||
| 17:08:28 | openstackgerrit | Chris Dent proposed openstack/nova master: Isolate placement database config https://review.openstack.org/541435 | |
| 17:08:28 | openstackgerrit | Chris Dent proposed openstack/nova master: WIP: Ensure that os-traits sync is attempted only at start of process https://review.openstack.org/553857 | |
| 17:09:12 | mnaser | kashyap: yeah so just to follow up, it looks like older libvirt versions have to do a whole lot of stuff on start up to get cpu capabilities | |
| 17:11:26 | openstackgerrit | Jim Rollenhagen proposed openstack/nova master: ironic: stop lying to the RT when ironic is down https://review.openstack.org/545479 | |
| 17:11:43 | jroll | this is what we were thinking, right? ^ | |
| 17:19:18 | superdan | fried_rolls: can you ack this again before I send it? https://review.openstack.org/#/c/552098/9 | |
| 17:22:00 | kashyap | mnaser: Yeah; noted. | |
| 17:25:10 | cdent | finucannot: do you mind if I unparent the db_driver conf removal change https://review.openstack.org/#/c/508487/5 it's parent looks stuck, but itself is indepdent | |
| 17:25:21 | finucannot | cdent: go for it | |
| 17:25:23 | openstackgerrit | melanie witt proposed openstack/nova master: only increment disk address unit for scsi devices https://review.openstack.org/538310 | |
| 17:25:40 | openstackgerrit | Chris Dent proposed openstack/nova master: conf: Remove 'db_driver' config opt https://review.openstack.org/508487 | |
| 17:25:45 | cdent | finucannot: thanks | |
| 17:26:24 | cdent | johnthetubaguy: you happy to put a +w on https://review.openstack.org/#/c/508487/ now that it has no parent? you +2'd it a while back. | |
| 17:33:36 | cdent | superdan: I'm not fried_rolls but looking through his comments, and checking the changes since last I looked, assuming we are good with the "followon" nits that eric mentioned on ps8 being followons, looks good to me | |
| 17:37:16 | superdan | cdent: yeah I just figured I'd give him another go at it, but if he doesn't show up soonish I can just send it | |
| 17:41:48 | superdan | mriedem: looks like it failed on glance again | |
| 17:41:51 | mnaser | is legacy-tempest-dsvm-cells failing more than usual? | |
| 17:42:02 | mriedem | mnaser: everything is failng more than usual | |
| 17:42:07 | mriedem | until the cinder revert merges | |
| 17:42:07 | mnaser | oh nice | |
| 17:42:09 | mnaser | its friday | |
| 17:42:17 | mnaser | okay, so ill hold rechecking stuff endlessly then | |
| 17:42:21 | mriedem | mnaser: https://review.openstack.org/#/c/553794/ | |
| 17:42:48 | openstackgerrit | Matt Riedemann proposed openstack/nova master: libvirt: handle DiskNotFound during update_available_resource https://review.openstack.org/553067 | |
| 17:42:59 | mriedem | superdan: well that's just not cool | |
| 17:43:18 | mriedem | superdan: oh that was the one i saw fail, pre-lunch | |
| 17:43:20 | mriedem | need to recheck | |
| 17:44:12 | superdan | oh I thought you already did | |
| 17:44:22 | superdan | heh I guess not, okay | |
| 17:44:37 | superdan | I was comparing your log link to the latest and like "yep, looks exactly the same" | |
| 17:44:42 | mriedem | the ovs multinode job, g-api started up right away on the subnode http://logs.openstack.org/08/553608/3/check/neutron-tempest-multinode-full/94a71ec/logs/subnode-2/devstacklog.txt.gz#_2018-03-16_14_40_43_011 | |
| 17:44:45 | mriedem | so not sure what's the difference | |
| 17:45:01 | superdan | unless LB networking means the nodes can't talk to each other | |
| 17:45:09 | mriedem | it might | |
| 17:45:27 | mriedem | https://review.openstack.org/#/c/553608/3/playbooks/legacy/neutron-tempest-linuxbridge/run.yaml@a33 | |
| 17:45:30 | mriedem | i can throw that back in there | |
| 17:45:35 | mriedem | but not sure if it will help | |
| 17:45:42 | superdan | I really thought we had ruled this out as an easy thing to do | |
| 17:45:53 | mriedem | well, the patch itself is pretty easy :) | |
| 17:45:54 | superdan | maybe I was just psychic | |
| 17:46:22 | mriedem | i.e. it's easy to copy the ovs multinode job, tweak for lb, and run it through ci until it works | |
| 17:46:27 | mriedem | not like i'm spending a lot of time on it | |
| 17:47:03 | mriedem | add PHYSICAL_NETWORK=default back in | |
| 17:47:05 | mriedem | *adding | |
| 17:55:07 | openstackgerrit | Matt Riedemann proposed openstack/nova master: api-ref: add a note in DELETE /os-services about deleting computes https://review.openstack.org/553598 | |
| 18:00:05 | jgwentworth | superdan, mriedem, leakypipes, fried_rolls, anyone else interested: here's a really rough draft on runways, please add comments and questions and if we can polish it a bit, I can send it out to the dev ML to start using the process https://etherpad.openstack.org/p/nova-runways-rocky | |
| 18:00:39 | jgwentworth | noting that this is going to be really not perfect and we just want to start with something. I should probably write that on the etherpad too | |
| 18:58:00 | fried_rolls | superdan: Gah, I thought I did that before I went to roll, sorry about that. Done. | |
| 18:58:21 | superdan | fried_rolls: thanks | |
| 19:02:18 | fried_rice | superdan, cdent: and by the way, here's my contribution: https://youtu.be/RYwgtJ7kJTI?t=130 | |
| 19:03:22 | superdan | fried_rice: now that's just silly | |
| 19:03:55 | fried_rice | Well, compared to the high cinema you presented, of course. | |
| 19:05:11 | fried_rice | superdan: I think the volume fails will continue to fail until https://review.openstack.org/#/c/553794/ merges? | |
| 19:05:43 | superdan | fried_rice: intermittently I think | |
| 19:06:03 | fried_rice | mm. Well, I was gonna wait to recheck mine, which is underneath his, so nyah. | |
| 19:06:25 | superdan | well, whatever, I was trying to help | |
| 19:06:32 | fried_rice | :) | |
| 19:06:39 | superdan | gonna go cry in my corner now | |
| 19:07:21 | fried_rice | What does that make the score? efried 1, dansmith 10000? | |
| 19:07:34 | fried_rice | I'm making my comeback | |
| 19:07:40 | mriedem | gah we don't have a Dict<?> field? | |
| 19:07:59 | superdan | mriedem: DictOfStrings | |
| 19:08:10 | mriedem | yeah but i've got a big ol bag of goodies | |
| 19:08:16 | mriedem | i know what to do, | |
| 19:08:17 | superdan | mriedem: a generic dict just makes it easy to throw complex types in and call it versioned | |
| 19:08:18 | mriedem | i'm just complaining | |
| 19:08:21 | mriedem | i know | |
| 19:08:31 | mriedem | bag-o-dicts | |
| 19:08:55 | mriedem | *space | |
| 19:09:01 | mriedem | my typing has been really terrible lately | |
| 19:09:06 | mriedem | talking to sean-k-mooney too much | |
| 19:21:14 | mriedem | same g-api failed to start error in that job http://logs.openstack.org/08/553608/4/check/neutron-tempest-linuxbridge/0a5acff/logs/subnode-2/devstacklog.txt.gz#_2018-03-16_19_15_53_511 | |
| 19:24:41 | mriedem | looking at devstack-gate, when it sets up multinode connectivity between the two hosts, it looks like it runs some ansible playbooks and those all assume ovs | |
| 19:25:27 | superdan | mriedem: assume how? | |
| 19:26:21 | mriedem | https://github.com/openstack-infra/devstack-gate/tree/master/playbooks/roles | |
| 19:29:20 | mriedem | although it looks like those are only used for nova-net multinode | |
| 19:29:24 | mriedem | or if using dvr | |
| 19:29:27 | mriedem | which this job isn't | |
| 19:38:08 | mriedem | the subnode code is trying to hit g-api on the primary node, not the subnode where g-api is also running | |
| 19:48:44 | superdan | aaaand, etherpad discussions fail. | |
| 19:49:40 | superdan | fried_rice: we might need a way to draw a cut line on a long series and say "this is what we expect to get done during this window".. it'd be nice if that was always a whole blueprint, but maybe that's not always realistic | |
| 19:50:05 | fried_rice | superdan: Yes, exactly so. | |
| 19:50:28 | fried_rice | For two reasons: 1) reviewer bandwidth, 2) reasonable expectation to have the whole series completely ready at the same time. | |
| 19:50:56 | fried_rice | There was a time when the upt series was over a dozen patches tall. | |
| 19:50:58 | superdan | entirely possible that early patches will be reviewed to oblivion making later patches irrelevant until substantially changed | |
| 19:51:05 | fried_rice | Yes, that also. | |
| 19:51:20 | superdan | well, the point of this is to maintain focus so we can chew through long series easier, | |
| 19:51:43 | superdan | so maybe it just needs to be "have approximately two weeks of stuff to review at the head of your series" | |
| 19:51:53 | superdan | nobody will ever be able to measure that of course | |
| 19:52:04 | fried_rice | No, but we'll get better at it with practice. | |
| 19:52:06 | superdan | and hopefully most blueprints will wholly fit into two weeks, but.. | |
| 19:52:10 | fried_rice | "Bottom N patches of bp X" | |
| 19:52:18 | fried_rice | kind of thing. | |
| 19:52:23 | mriedem | it's likely a discussion that happens when talking about the next thing to go into a slot, | |
| 19:52:27 | superdan | well, aim for whole blueprints, but for special problem children like you... | |
| 19:52:31 | mriedem | for single-patch small bp's, it's not an issue | |
| 19:52:56 | superdan | yeah | |
| 19:53:32 | mriedem | "is bp x ready for a slot? if so, are you going to be around for the next 2 weeks to respond to reviews? if so, i see that it's 20 patches already, what is the reasonable target?" | |
| 19:54:38 | mriedem | this shouldn't be all that process heavy | |
| 19:54:51 | mriedem | says the guy that loves process | |