| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-09-18 | |||
| 17:28:08 | dtantsur | np | |
| 17:28:28 | dtantsur | mriedem: we also have a traits discussion with johnthetubaguy in our room, so you don't have the whole context for sure | |
| 17:28:32 | dtantsur | s/have/had/ | |
| 17:38:17 | cfriesen | mriedem (or anyone): I heard something about a "prioritization discussion" from last tuesday that pushed shared resources down in priority...is there an etherpad/link for that? | |
| 18:01:47 | leitan_ | is nova caching the rabbit connection string somewhere ? | |
| 18:10:35 | mriedem | leitan_: yeah, check the nova_api.cell_mappings table | |
| 18:10:44 | mriedem | and there is a context cache in the service | |
| 18:11:19 | leitan_ | mriedem: holy molly thanks ! | |
| 18:11:33 | mriedem | leitan_: #2 here https://docs.openstack.org/nova/latest/user/cells.html#faqs | |
| 18:12:16 | mriedem | cfriesen: not much details but it's at the bottom here https://etherpad.openstack.org/p/nova-ptg-queens-placement | |
| 18:12:37 | leitan_ | mriedem: yes, now i see that | |
| 18:12:45 | leitan_ | it should be there where i persisted the rabbit | |
| 18:12:55 | leitan_ | without ssl | |
| 18:13:25 | leitan_ | and when i change the config ... i forgot updating the cell | |
| 18:14:04 | mriedem | leitan_: there is an update_cell CLI https://docs.openstack.org/nova/latest/cli/nova-manage.html#nova-cells-v2 | |
| 18:14:47 | leitan_ | mriedem: thanks a lot ! | |
| 18:15:16 | mriedem | yw | |
| 18:15:59 | mriedem | mikal: just a heads up, i don't plan on sending a ptg recap on the privsep stuff. feel free if you want. | |
| 18:26:59 | sdague | he's still in a mountain somewhere | |
| 18:39:45 | dansmith | hopefully on a mountain and not in it | |
| 18:53:55 | avolkov | mriedem: yes, will try to work on osc for placement, thanks for the info | |
| 19:37:23 | mriedem | ew https://bugs.launchpad.net/nova/+bug/1717916 | |
| 19:37:25 | openstack | Launchpad bug 1717916 in OpenStack Compute (nova) "Nova won't reschedule when specific hypervisor is set and request failed" [Undecided,New] | |
| 19:38:24 | mriedem | "if len(force_hosts) == 1 and (len(nodes) == 1 and len(force_nodes) == 1) or not force_nodes" | |
| 19:41:01 | sdague | mriedem: I'm confused why that's a new issue | |
| 19:41:15 | mriedem | we don't reschedule across nodes for the same host | |
| 19:41:27 | mriedem | so, create a server with the AZ:HOST thing, | |
| 19:41:29 | mriedem | but no NODE | |
| 19:41:32 | sdague | ok | |
| 19:41:37 | mriedem | if it's ironic, you could fail the first node and we won't retry | |
| 19:41:40 | sdague | but, that test has existed for over a year | |
| 19:41:55 | sdague | that's how we test that multinode is actually doing that | |
| 19:44:05 | mriedem | i don't know why it's a recent problem for ironic, unless they've been blacklisting that test this whole time | |
| 19:45:08 | sdague | yeh, I feel like that is the specific ask back. Because to my knowledge that test has been pretty static for a while. | |
| 19:45:35 | sdague | and this might be the team expanding CI and not running long enough before voting on something to realize this was going to happen | |
| 19:46:13 | mriedem | yeah just asked for details in the bug report | |
| 19:49:28 | mriedem | dansmith: look familiar at all? https://bugs.launchpad.net/nova/+bug/1717915 | |
| 19:49:29 | openstack | Launchpad bug 1717915 in OpenStack Compute (nova) "nova services and transport_url, cannot connect to vhost if specified" [Undecided,New] | |
| 19:50:44 | dansmith | mriedem: well, no, but I think he broke his config | |
| 19:50:46 | dansmith | I'll comment | |
| 19:50:58 | mriedem | i just did too, transport_url isn't a list opt | |
| 19:51:30 | dansmith | mriedem: I think it is supposed to take a list of hosts for clustered rabbit, I don't know if that's the problem | |
| 19:52:14 | mriedem | ah the docs aren't clear on the options | |
| 19:52:21 | mriedem | rabbit_hosts is deprecated in favor of transport_url | |
| 19:52:42 | dansmith | right | |
| 19:53:26 | dansmith | so yeah, I think this is just him getting the right format for transport_url with multiple hosts | |
| 19:53:44 | dansmith | i.e. hopefully not a cells issue but just a transport_url one | |
| 19:55:50 | mriedem | heh yeah, because we aren't parsing that in nova as a list | |
| 19:57:06 | mriedem | from the oslo.messaging code, it looks like it splits on ; | |
| 19:58:27 | mriedem | no tests on the change that deprecated the option... | |
| 20:01:12 | mriedem | https://github.com/openstack/oslo.messaging/commit/2f0d53b | |
| 20:02:48 | dansmith | well they'd be tests on the transport_url change right? | |
| 20:10:52 | mriedem | yeah, and looks like it does split on comma on the parsed netloc https://github.com/openstack/oslo.messaging/blob/393ecff3451091404832dd6b8a088e1bec760101/oslo_messaging/transport.py#L512 | |
| 20:12:25 | mriedem | and i think it applies the virtual host to all of them | |
| 20:37:40 | mriedem | rhosp is clustered rabbit right? so i'd assume you guys would have hit this with cells v2 by now if it were an obvious issue | |
| 20:38:48 | jaypipes | mriedem: tyvm for doing the denver recaps. very useful. | |
| 20:39:11 | jaypipes | mriedem: and no, I don't think rhosp is clustered rabbit, unless I'm mistaken... dansmith? | |
| 20:39:33 | dansmith | it absolutely is | |
| 20:40:17 | dansmith | mriedem: tripleo HA does it as well, so we can probably just ask a tripleo person, although the one I'd ask is on PTO this week, AFAK | |
| 20:40:20 | dansmith | *AFAIK | |
| 20:41:05 | mriedem | rhosp is tripled everything | |
| 20:41:08 | mriedem | even the deployer | |
| 20:41:26 | dansmith | owalsh likely knows, but he's probably not around this late | |
| 20:41:27 | mriedem | it's like the ihop breakfast of openstack deployments | |
| 20:41:28 | dansmith | in general, you can assume owalsh knows everything about this stuff | |
| 20:41:36 | dansmith | he lazy-loads things he doesn't know off the top of his head | |
| 20:42:57 | mriedem | no props for my 2 x 2 x 2 reference? | |
| 20:44:03 | mriedem | jaypipes: thanks. prepare for the "everything else" finale. | |
| 20:44:42 | dansmith | mriedem: too obscure for me, but: haha | |
| 20:44:52 | jaypipes | dansmith: ah, ok, sorry. for some reason I thought triple-o didn't configure rabbitmq in cluster mode.. | |
| 20:45:23 | dansmith | unless something has changed, but I don't think so :) | |
| 20:45:40 | jaypipes | dansmith: no, prolly just me badly misremembering something | |
| 21:06:21 | mriedem | man, the notes on preemptible instances are confusing | |
| 21:06:29 | mriedem | 'first phase, don't touch quota' | |
| 21:06:36 | mriedem | 'the reaper will modify quota' | |
| 21:53:31 | efried | alex_xu Should be back in a couple hours :) | |
| 21:55:28 | clarkb | trivia time, the size of the nova repo and number of changes to the nova repo and total number of comments on nova all have a direct impact on the length of the downtime :) (but don't let that stop you from using gerrit, its just annoying gerrit has upgrade scaling issues with the size of projects) | |
| 22:24:05 | jaypipes | clarkb: you know there's a full copy of OpenWRT embedded in the Nova source tree, right? happy trivia day :) | |
| 22:24:35 | clarkb | I did know that :) | |
| 22:24:42 | clarkb | if you ever need a vm image ot boot its there | |
| 22:37:32 | dims | jaypipes : seriously? wow | |
| 22:57:41 | openstackgerrit | Nicolas Simonds proposed openstack/nova master: libvirt: add support for virtio-net rx/tx queue sizes https://review.openstack.org/484997 | |
| 22:59:40 | jaypipes | dims: yup :) | |
| 23:01:22 | dansmith | mriedem_afk: daddy don't live in that new york city no more | |
| 23:45:22 | dansmith | ohai gerrit | |
| 23:55:44 | mriedem_afk | dansmith: song reference? | |
| 23:56:07 | mriedem | or you just referring to someone as "daddy"? | |
| #openstack-nova - 2017-09-19 | |||
| 00:01:38 | dansmith | mriedem: t'was playing and thought of you (because artist, not because....any other reasons) | |
| 00:11:05 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Account for compute.metrics.update in legacy notification whitelist https://review.openstack.org/504978 | |
| 00:13:56 | openstackgerrit | Matt Riedemann proposed openstack/nova stable/pike: Functional test for regression bug #1713783 https://review.openstack.org/504943 | |
| 00:13:58 | openstack | bug 1713783 in OpenStack Compute (nova) pike "After failed evacuation the recovered source compute tries to delete the instance" [High,Triaged] https://launchpad.net/bugs/1713783 | |
| 00:13:58 | openstackgerrit | Matt Riedemann proposed openstack/nova stable/pike: Set error state after failed evacuation https://review.openstack.org/504979 | |
| 00:16:48 | openstackgerrit | Merged openstack/nova master: [placement] Unregister the ResourceClassList object https://review.openstack.org/502154 | |
| 01:19:48 | openstackgerrit | Naichuan Sun proposed openstack/nova master: xenapi: Support live migration in pooled multi-nodes environment https://review.openstack.org/489451 | |
| 01:39:41 | openstackgerrit | Michael Still proposed openstack/nova master: Use symbolic names for capabilities, expand sys_admin context. https://review.openstack.org/504193 | |
| 01:39:41 | openstackgerrit | Michael Still proposed openstack/nova master: Move the dac_admin privsep code to a new location. https://review.openstack.org/504194 | |
| 01:39:42 | openstackgerrit | Michael Still proposed openstack/nova master: Squash dac_admin privsep context. https://review.openstack.org/504195 | |
| 01:39:42 | openstackgerrit | Michael Still proposed openstack/nova master: Squash dacnet_admin privsep context. https://review.openstack.org/504805 | |
| 01:39:43 | openstackgerrit | Michael Still proposed openstack/nova master: Move libvirts dmcrypt support to privsep. https://review.openstack.org/490737 | |
| 01:39:43 | openstackgerrit | Michael Still proposed openstack/nova master: Move ploop commands to privsep. https://review.openstack.org/492325 | |