| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-02-01 | |||
| 19:33:59 | vsaienko | mriedem yes | |
| 19:37:57 | johnsom | So, looking at that trace, I'm not sure if it's a nova, neutron, or oslo issue... | |
| 19:38:39 | mriedem | dansmith: want me to work https://review.openstack.org/#/c/426826/ into https://review.openstack.org/#/c/424737/ or should we just mention the periodic task in the cells v2 devref? | |
| 19:38:53 | manjeets | thanks mriedem | |
| 19:39:17 | mriedem | or you couldn't care less and wish i'd stop asking about docs | |
| 19:41:04 | ildikov | mriedem: sdague: hi | |
| 19:41:31 | mriedem | local.conf for cinder v3 testing right? | |
| 19:41:48 | ildikov | mriedem: sdague: just wanted to ask what would be the best timing to get to the details of figuring out the Cinder V3 testing in the placement API job? | |
| 19:42:05 | dansmith | mriedem: probably needs to be in both places | |
| 19:42:08 | dansmith | mriedem: I'll do it | |
| 19:42:29 | mriedem | dansmith: ok, was just adding the "Once fully upgraded" to the first issue bullet | |
| 19:42:52 | dansmith | mriedem: oh I can do that too, I just got distracted with real stuff | |
| 19:43:21 | dansmith | but whatever you ant | |
| 19:43:22 | dansmith | *want | |
| 19:43:36 | mriedem | dansmith: or, just throw the mention of the periodic into a new patch on top of https://review.openstack.org/#/c/424737/ with it's own reno | |
| 19:43:41 | mriedem | i don't care too much either way | |
| 19:43:52 | mriedem | ildikov: probably next week | |
| 19:44:10 | dansmith | mriedem: the periodic needs a recheck anyway, so if you drop the -W I can just add it in there | |
| 19:44:11 | mriedem | ildikov: it's going to require being able to pass a post-config via local.conf from the job definition in project-config through d-g into devstack | |
| 19:44:18 | dansmith | er, drop the +W | |
| 19:44:38 | mriedem | dansmith: done | |
| 19:45:05 | ildikov | mriedem: we are confident at this point that the post-config direction will be accepted, right? | |
| 19:45:18 | ildikov | mriedem: I guess we don't have too many choices here anyhow though | |
| 19:45:59 | mriedem | ildikov: sdague is the one making it happen i believe | |
| 19:46:05 | mriedem | because it's the direction he wants | |
| 19:46:18 | mriedem | so yes i'm pretty confident | |
| 19:46:37 | ildikov | mriedem: great, that sounds like a good starting point | |
| 19:46:39 | ildikov | thanks | |
| 19:48:16 | sdague | yeh, the biggest issue is we're now past freeze so until everyone is on an open master again, poking at devstack-gate in that way is going to be dangerous | |
| 19:48:49 | sdague | ildikov: https://review.openstack.org/#/c/326585/ was the change that got the base support over | |
| 19:50:26 | mriedem | f yes http://logs.openstack.org/26/426926/6/check/gate-grenade-dsvm-neutron-ubuntu-xenial/9019244/logs/grenade.sh.txt.gz#_2017-02-01_19_01_54_881 | |
| 19:50:38 | mriedem | dansmith: jaypipes: melwitt: sdague: bauzas: ^ nova-status upgrade check in grenade | |
| 19:51:03 | sdague | mriedem: cool | |
| 19:51:49 | openstackgerrit | Dan Smith proposed openstack/nova: Allow scheduler to run cell host discovery periodically https://review.openstack.org/426826 | |
| 19:51:58 | ildikov | sdague: got it, thanks! | |
| 19:52:09 | dansmith | mriedem: ^ | |
| 19:52:13 | mriedem | yup readibg | |
| 19:52:40 | dansmith | mriedem: that status check gives me butterflies | |
| 19:52:43 | ildikov | sdague: so basically as soon as master gets open for Pike we can get it sorted out | |
| 19:54:23 | mriedem | dansmith: one minor thing in https://review.openstack.org/#/c/426826/6/releasenotes/notes/scheduler-can-discover-hosts-4b799cbd14dbc7dc.yaml | |
| 19:54:33 | dansmith | ah yeah | |
| 19:55:02 | openstackgerrit | Dan Smith proposed openstack/nova: Allow scheduler to run cell host discovery periodically https://review.openstack.org/426826 | |
| 19:55:03 | sdague | ildikov: yeh, it's going to take some time to work through things. Honestly, that would probably be worth PTG time on the infra / qa side to sketch out the path forward | |
| 19:55:21 | sdague | because it's not a closed loop yet | |
| 19:55:22 | mriedem | ildikov: i think infra has an etherpad for ptg | |
| 19:55:28 | mriedem | ildikov: probably worth getting that on there | |
| 19:55:40 | sdague | and honestly been having a heck of a time getting that devstack-gate patch reviewed | |
| 19:55:41 | mriedem | https://etherpad.openstack.org/p/infra-ptg-pike | |
| 19:56:26 | mriedem | jaypipes: meet me in -operators | |
| 19:59:26 | mriedem | dansmith: is 0 positive? | |
| 19:59:37 | mriedem | dansmith: or did you not want to distinguish between <0 for the periodic? | |
| 19:59:40 | mriedem | since 0 will make it run | |
| 19:59:48 | dansmith | mriedem: zero is positive | |
| 20:01:11 | mriedem | sfinucan: https://review.openstack.org/#/c/426826/7 | |
| 20:01:18 | mriedem | sfinucan: latest adds a release note | |
| 20:03:08 | mriedem | dansmith: if you have butterflies about that nova-status + grenade thing, then https://review.openstack.org/#/c/427499/ is up your alley | |
| 20:03:29 | mriedem | https://review.openstack.org/#/c/426926/ was the grenade change testing it | |
| 20:10:10 | openstackgerrit | Matt Riedemann proposed openstack/nova: doc: add some detail to the map_cell0 man page https://review.openstack.org/427450 | |
| 20:10:10 | openstackgerrit | Matt Riedemann proposed openstack/nova: doc: refer to the cell_v2 man pages from the cells v2 doc https://review.openstack.org/427459 | |
| 20:10:11 | openstackgerrit | Matt Riedemann proposed openstack/nova: doc: update the man page entry for nova-manage db sync https://review.openstack.org/427460 | |
| 20:10:52 | ildikov | mriedem: sdague: thanks, I added the patch and a note to the etherpad | |
| 20:21:19 | artom | http://logs.openstack.org/59/424759/12/gate/gate-nova-releasenotes/d0bf975/console.html#_2017-02-01_19_58_38_712633 | |
| 20:21:21 | artom | *stab* | |
| 20:21:30 | artom | *stab stab stab stab* | |
| 20:21:51 | artom | (Unrelated to any important currently ongoing issues) | |
| 20:22:15 | mriedem | artom: oh that sucks | |
| 20:22:17 | mriedem | 3 in the gate | |
| 20:23:32 | artom | The gate is Kanye'ing me | |
| 20:23:53 | artom | "Your patch is nice and imma let in finish - but this guy - this guy has the best patch of all time!" | |
| 20:25:17 | openstackgerrit | Dan Smith proposed openstack/nova: Allow scheduler to run cell host discovery periodically https://review.openstack.org/426826 | |
| 20:25:18 | openstackgerrit | Dan Smith proposed openstack/nova: Reserve migration placeholders for Ocata backports https://review.openstack.org/427900 | |
| 20:25:18 | openstackgerrit | Dan Smith proposed openstack/nova: Add ComputeNode.mapped field https://review.openstack.org/427901 | |
| 20:25:19 | openstackgerrit | Dan Smith proposed openstack/nova: Add ability to query for ComputeNodes by their mapped value https://review.openstack.org/427902 | |
| 20:25:19 | openstackgerrit | Dan Smith proposed openstack/nova: Make discover_hosts only query for unmapped ComputeNode records https://review.openstack.org/427903 | |
| 20:26:06 | mriedem | bauzas: replied to your comments in https://review.openstack.org/#/c/427367/ but i didn't see anything that i needed to change | |
| 20:26:13 | mriedem | i might be misunderstanding your points though | |
| 20:26:23 | dansmith | mriedem: ^ is what I'm thinking, sans any tests or testing | |
| 20:26:43 | mriedem | tests i can do without, but not testing | |
| 20:26:44 | mriedem | :) | |
| 20:27:08 | dansmith | yeah, well, eventually, I just wanted it out there | |
| 20:27:21 | dansmith | in case I stab a fork in my eye between now and pike | |
| 20:27:22 | dansmith | also, the migration backports are in that stack | |
| 20:27:27 | dansmith | placeholders I mean | |
| 20:27:34 | bauzas | mriedem: sorry, I meant in https://review.openstack.org/#/c/427367/2/doc/source/placement.rst@175 that a scheduler hard failure could also happen once you upgrade all your computes, not only in Pike | |
| 20:28:14 | mriedem | bauzas: oh i see | |
| 20:29:09 | bauzas | mriedem: and about the bullet good place, I meant that since operators need to modify their nova.conf for adding the placement section, it could be nice to let them know quite close that they also need to update ratios | |
| 20:29:17 | bauzas | but that's a nit | |
| 20:29:23 | mriedem | bauzas: even if n-sch can't talk to placement after you've upgraded your computes, | |
| 20:29:35 | mriedem | we'll fallback to querying the compute_nodes table | |
| 20:29:42 | mriedem | so it's not the same kind of hard failure | |
| 20:29:44 | bauzas | mriedem: not exactly | |
| 20:30:01 | bauzas | mriedem: because if you don't update the n-sch nova.conf, then it would return NoValidHosts | |
| 20:30:02 | mriedem | none of the compute services would be running, which would be a different kind of failure | |
| 20:30:25 | johnsom | I have opened a bug for the os_vif issue I am seeing: https://bugs.launchpad.net/nova/+bug/1661086 | |
| 20:30:25 | openstack | johnsom: Error: Could not gather data from Launchpad for bug #1661086 (https://launchpad.net/bugs/1661086). The error has been logged | |
| 20:30:59 | johnsom | Ha, everything is melting down... | |
| 20:31:24 | bauzas | mriedem: I mean, if the operator modifies nova.conf for all computes but forgets to do so for n-sch, and then upgrades all their computes, then we have the scheduler calling Placement, righT? | |
| 20:31:25 | mriedem | bauzas: oh i see this https://github.com/openstack/nova/blob/master/nova/scheduler/filter_scheduler.py#L184 | |
| 20:31:38 | bauzas | mriedem: exactly | |
| 20:32:00 | bauzas | that's what I call an hard failure | |