| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-09-01 | |||
| 13:32:43 | mriedem | is it changing the response? | |
| 13:32:52 | mriedem | if it's changing the response then it's a new microversion | |
| 13:32:52 | figleaf | yes | |
| 13:32:56 | cdent | sdague is probably the main arbiter on this kind of thing | |
| 13:33:12 | cdent | mriedem: the squiggle is that we changed the content of that section before without a microversion | |
| 13:33:22 | cdent | thus the lawyering going on now | |
| 13:33:36 | figleaf | two clouds requesting the same microversion would get different data | |
| 13:33:42 | mriedem | past performance is not an indicator of future earnings | |
| 13:34:13 | mriedem | what was the previous non-versioned change? | |
| 13:34:41 | cdent | Yeah, the rules are it must be a new microversion. The reason I’m being squishy about it is just to reinforce that just because I wrote the latest version of the rules doesn’t mean I’m in love with microversions and want to marry them. | |
| 13:34:52 | fried_rice | mriedem https://review.openstack.org/#/c/468923/1/nova/api/openstack/placement/handlers/resource_provider.py | |
| 13:34:53 | cdent | mriedem: it’s linked in fried_rice’s comments | |
| 13:34:54 | cdent | jinx | |
| 13:35:15 | mriedem | let me note that it's too early in my day to have to deal with the age old, "does this require a microversion?" question | |
| 13:35:49 | figleaf | it's the usual problem: a seemingly backwards-compatible change | |
| 13:36:13 | mriedem | so alex_xu and sdague approved https://review.openstack.org/#/c/468923/ | |
| 13:36:20 | mriedem | with no comment on the microversion part of this | |
| 13:36:29 | fried_rice | right | |
| 13:38:10 | cdent | fried_rice: did you find this because you are actually using the links, or just because you were having a poke around. If someone’s using genuine hypertextuality, i’m going to need to sit down. | |
| 13:39:10 | cdent | (somewhat related, I’ve started using genuine hypertextuality in my gabbi tests of placement, just for the sake of it) | |
| 13:39:20 | fried_rice | cdent Poking around. Larnin' up the placement API and resource providers and whatnot. | |
| 13:39:32 | fried_rice | Working towards being able to do generic device management via placement. | |
| 13:39:46 | cdent | Larnin’!!!??? Ain’t no body got time for that. | |
| 13:40:16 | fried_rice | The alternative is putting together a huge pile of incomprehensible sh*t that will waste everybody's time in Denver. | |
| 13:40:36 | fried_rice | This way it gets to be a huge pile of slightly-comprehensible sh*t. | |
| 13:41:19 | kashyap | Huh, the documentation for 'live_migration_scheme' says: "Override the default libvirt live migration scheme", but it doesn't mention _what_ are the possible values? | |
| 13:41:23 | kashyap | From here: https://docs.openstack.org/ocata/config-reference/compute/config-options.html | |
| 13:41:30 | cdent | I hope to introduce you to some vmware folk who are curious about that stuff too, and are also planning to do some funky not-just-for-the-sake-of-nova resource mgt via placement | |
| 13:41:55 | cdent | s/you/fried_rice/ | |
| 13:42:14 | mriedem | kashyap: sec, i have a ML thread link | |
| 13:42:37 | kashyap | mriedem: Ah, I came across it (as I sometimes filter openstack-dev list by your e-mail ID :P) | |
| 13:42:38 | fried_rice | cdent Cool, I've been wanting to talk to VMWare about this. | |
| 13:42:40 | kashyap | Let me go look | |
| 13:42:55 | fried_rice | Cause I think they have some of the same issues as claudiub and I do. | |
| 13:43:09 | fried_rice | that we want to make sure the new solution solves. | |
| 13:43:31 | cdent | fried_rice: yup | |
| 13:43:57 | mriedem | kashyap: http://lists.openstack.org/pipermail/openstack-dev/2017-June/117987.html | |
| 13:44:46 | kashyap | mriedem: BTW, thanks for kickstarting the cleanup of the spaghetti that is LM doc - 9c2518c2 ("Cleanup the plethora of libvirt live migration options") | |
| 13:45:47 | mriedem | yeah it's impossible | |
| 13:46:48 | kashyap | Heh, "landmine of terrible" | |
| 13:47:12 | openstackgerrit | Elod Illes proposed openstack/nova master: Functional test for regression bug #1713783 https://review.openstack.org/500057 | |
| 13:47:13 | 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 | |
| 13:50:00 | kashyap | mriedem: Darn, I missed your e-mail as it got filtered to the 'operators' mail dir. | |
| 13:55:25 | mriedem | elod: thanks for https://review.openstack.org/#/c/500057/ - just a few small things inline | |
| 13:57:55 | elod | mriedem: thanks too, checking | |
| 13:59:03 | kashyap | Although you could figure out for context, but the URL will be useless in future | |
| 13:59:08 | kashyap | s/for/from/ | |
| 14:01:42 | gibi | kashyap: you can have a github link that points to place via a specfic commit hash. That is futureproof | |
| 14:07:07 | mriedem | e.g. https://github.com/openstack/nova/blob/7815108d4892525b0047c787cbd2fe2f26c204c2/nova/virt/libvirt/driver.py#L652 | |
| 14:07:10 | mriedem | like in my email | |
| 14:07:53 | stephenfin | Any chance of a +W on this? Been sitting around for a while https://review.openstack.org/#/c/407055/ | |
| 14:08:19 | kashyap | gibi: Ah, noted | |
| 14:08:35 | mriedem | stephenfin: the change under it has a -1 on it | |
| 14:08:39 | mriedem | so what's the rush | |
| 14:08:42 | mriedem | unless you re-order | |
| 14:08:49 | kashyap | Just reading this one (oh dear, yet more live migration options) - https://review.openstack.org/#/c/456571/ | |
| 14:08:50 | stephenfin | Aha, correct. nvm me | |
| 14:09:08 | stephenfin | My list of open patches has grown rather large/unmanageable :/ | |
| 14:09:19 | kashyap | I know it intends to solve the said valid bug where one can't specify one of the URI variants | |
| 14:09:24 | mriedem | kashyap: the point of that is to remove the URI option and make it specific | |
| 14:09:29 | mriedem | so you can configure the parts of the URI | |
| 14:09:32 | mriedem | right | |
| 14:09:51 | mriedem | live_migration_query_parameters :) | |
| 14:09:58 | kashyap | mriedem: :-) I see the point | |
| 14:10:02 | mriedem | oh i guess that's extra_params | |
| 14:10:06 | mriedem | i was just joking about that one | |
| 14:10:46 | kashyap | Yeah, I've read your comment where you lament about the proliferation of the options | |
| 14:11:16 | mriedem | you know, it makes me think, if we merged that and removed live_migration_uri, in 2 years time someone will say, "why do i have to configure separately the 4 parts of this URI? why can't i just specify the single live migration URI as a config option?" | |
| 14:11:28 | mriedem | that someone will likely be me | |
| 14:12:32 | mriedem | oh i guess it's because only nova knows the host | |
| 14:12:38 | mriedem | so we have to sub in the host/ip in the URI | |
| 14:12:47 | kashyap | On a call, /me is a bit slow to read the scroll here | |
| 14:13:00 | kashyap | mriedem: Exactly! About the future complainer | |
| 14:13:24 | kashyap | But yeah, it's a bit hair-splitty | |
| 14:17:22 | openstackgerrit | Elod Illes proposed openstack/nova master: Functional test for regression bug #1713783 https://review.openstack.org/500057 | |
| 14:17:23 | 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 | |
| 14:17:53 | mriedem | gibi: we might want to think about how to highlight http://docs-draft.openstack.org/63/499963/1/check/gate-nova-docs-ubuntu-xenial/2c3bf74//doc/build/html/reference/notifications.html#existing-versioned-notifications at a higher level in the docs, | |
| 14:18:08 | mriedem | gibi: because right now you have to go main index > technical reference deep dives > notifications > samples | |
| 14:18:43 | mriedem | if i'm a consumer of the versioned notifications and want samples, like we have for the compute api reference, it would be nice to get to those from a higher level since i dont think i'd expect to find it in 'technical reference deep dives' | |
| 14:19:13 | gibi | mriedem: sure, I have to scan the new structure to see where a link can fit | |
| 14:19:24 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Remove dest node allocation if evacuate MoveClaim fails https://review.openstack.org/499878 | |
| 14:19:25 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Add recreate test for evacuate claim failure https://review.openstack.org/499874 | |
| 14:19:25 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Add a test to make sure failed evacuate cleans up dest allocation https://review.openstack.org/499877 | |
| 14:20:08 | mriedem | gibi: maybe a new section in https://docs.openstack.org/nova/latest/#for-end-users ? | |
| 14:20:23 | mriedem | because the tech reference deep dive is under 'For Contributors' | |
| 14:20:33 | mriedem | but if i'm a notification consumer i'm not necessarily a contributor | |
| 14:20:47 | mriedem | i'll hack something up quick to fight over | |
| 14:21:15 | gibi | mriedem: I think similar to what is in https://docs.openstack.org/nova/latest/#writing-to-the-api | |
| 14:21:20 | openstackgerrit | Eric Fried proposed openstack/nova master: Include /resource_providers/uuid/allocations link https://review.openstack.org/499826 | |
| 14:21:31 | mriedem | yeah | |
| 14:21:32 | gibi | mriedem: or maybe even inside it as a sub section | |
| 14:21:39 | gibi | mriedem: as notification is also an API | |
| 14:22:32 | mriedem | ok i can stack that on top of this https://review.openstack.org/#/c/498977/ | |
| 14:23:07 | gibi | mriedem: cool. but if you have more important things then I can do that doc change | |
| 14:23:28 | mriedem | this is the most important thing in my life right this second | |
| 14:23:39 | gibi | mriedem: thank go for it :) | |
| 14:24:42 | gibi | I mean thanks, then go for it | |
| 14:27:18 | openstackgerrit | Eric Fried proposed openstack/nova master: Glance download: only fsync files https://review.openstack.org/451094 | |
| 14:27:49 | openstackgerrit | Stephen Finucane proposed openstack/nova master: conf: Remove '[conductor] topic' opt https://review.openstack.org/499179 | |
| 14:28:46 | mriedem | elod: can you rebase https://review.openstack.org/#/c/499237/ on top of https://review.openstack.org/#/c/500057/ now? | |