| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-08-02 | |||
| 20:04:37 | openstackgerrit | Chris Dent proposed openstack/nova master: Update RT aggregate map less frequently https://review.openstack.org/489633 | |
| 20:05:02 | mriedem | fixed in 2.47, | |
| 20:05:09 | mriedem | but this is like mitaka -> pike | |
| 20:05:23 | mriedem | good times | |
| 20:05:27 | sdague | mriedem: right, that was the crux of the fight around whether the original flavor id was included in the embedded flavor structure | |
| 20:05:35 | sdague | because of that feature in horizon | |
| 20:05:49 | mriedem | we should have added a cell0 for flavors :) | |
| 20:05:54 | mriedem | nova_flavors db | |
| 20:06:04 | sdague | which lets people dig themself a hole to fall in after they shoot themselves in the foot | |
| 20:06:05 | mriedem | when you just can't get enough dbs | |
| 20:06:21 | sdague | I'm telling you, db per project | |
| 20:06:27 | mriedem | sounds like a fun game | |
| 20:10:10 | openstackgerrit | Rawan Herzallah proposed openstack/nova master: Adding NVMEoF for libvirt driver https://review.openstack.org/482640 | |
| 20:12:10 | mriedem | so i think Kevin_Zheng is going to send something to the ML asking about changing that behavior in horizon | |
| 20:12:29 | dansmith | meaning removing flavor editing? | |
| 20:12:37 | openstack | Launchpad bug 1708260 in OpenStack Compute (nova) "Sending empty allocations list on a PUT /allocations/{consumer_uuid} results in 500" [Medium,Confirmed] - Assigned to Chris Dent (cdent) | |
| 20:12:37 | cdent | dansmith, jaypipes, edleafe: https://bugs.launchpad.net/nova/+bug/1708260 you want to provide an opinion on whether the response should be a 400 or a lukewarm success (you asked me to do nothing, I have successfully done nothing) | |
| 20:12:46 | mriedem | at least disabling the ability to do it on the instance record itself, | |
| 20:13:03 | mriedem | like, there is a panel showing the instance and it's flavor and the ability to edit it right there on the instance record | |
| 20:13:09 | mriedem | which we know doesn't actually resize the instance or anythign | |
| 20:13:22 | dansmith | cdent: 200 vs 400 seems like something I'd have an uninformed gut opinion on, which you would immediately whip out some document to refute | |
| 20:13:26 | dansmith | cdent: so... no, I don't care :) | |
| 20:13:30 | dansmith | cdent: 500 seems wrong | |
| 20:13:55 | cdent | dansmith: aw dan, I’m trying to be inclusive. | |
| 20:14:06 | mriedem | 409 | |
| 20:14:09 | mriedem | always 409 | |
| 20:14:35 | dansmith | cdent: I'm being tongue-in-cheeky | |
| 20:14:44 | cdent | me too | |
| 20:15:09 | cdent | I’m going to call is a schema violation for now | |
| 20:15:35 | openstackgerrit | Jackie Truong proposed openstack/nova master: Add trusted_certs to Instance object https://review.openstack.org/489408 | |
| 20:15:43 | edleafe | If the allocations body is required, 409 seems like the correct response | |
| 20:16:07 | mriedem | ha | |
| 20:16:10 | mriedem | wouldn't it be a 400? | |
| 20:16:42 | cdent | that is, like, so wrong | |
| 20:16:47 | cdent | it would be 400 | |
| 20:16:54 | mriedem | 402? | |
| 20:16:59 | mriedem | i'll take your nothing and charge you for it | |
| 20:17:08 | mriedem | haha | |
| 20:17:28 | edleafe | sorry, fat finger | |
| 20:17:32 | edleafe | yes, 400 | |
| 20:17:41 | dansmith | i'll hit the brakes and he'll fly right by | |
| 20:18:21 | sdague | 418 and call it a day :) | |
| 20:18:23 | cdent | so anyway, edleafe: you’ve hit crux of my query: is it required? | |
| 20:18:41 | cdent | one could argue that sending an empty allocations list is some kind of delete | |
| 20:18:45 | cdent | but that just feels icky | |
| 20:18:51 | cdent | so I’m going to change the schema | |
| 20:19:15 | sdague | cdent: 400 on that and call it a schema violation seems right to me | |
| 20:19:37 | edleafe | cdent: Having an allocation list is required. Populating it with allocations is not | |
| 20:19:45 | sdague | the point of doing that is to help the user write a better application by erroring on them when they might have typoed a thing | |
| 20:19:46 | edleafe | So yeah, an empty list == delete | |
| 20:19:55 | dansmith | mriedem: come on I get no love for the 80s movie reference? | |
| 20:20:33 | sdague | edleafe: a delete should be explicit, and not mistaken for a forgotten initialization | |
| 20:21:12 | edleafe | sdague: agree. So a schema change seems like the best solution | |
| 20:21:18 | mriedem | dansmith: was coding | |
| 20:21:26 | mriedem | i see | |
| 20:21:26 | mriedem | oh top gun | |
| 20:21:28 | mriedem | good one mav | |
| 20:22:56 | cdent | edleafe, sdague: it is likely that before we added the consumer_id to the object, it operated as an unintentional delete | |
| 20:23:08 | dansmith | by the way, I'll be taking up a collection to buy one of these for mriedem: https://chummytees.com/products/i-speak-fluent-movie-quotes-t-shirt-hoodie-tank-top | |
| 20:24:05 | cfriesen | I was waiting for "she's real fine my 409" | |
| 20:24:08 | melwitt | he can wear it when he gives summit presentations | |
| 20:25:30 | mriedem | cfriesen: isn't that beach boys ala 60s? | |
| 20:25:34 | mriedem | too old | |
| 20:25:37 | openstackgerrit | Chris Dent proposed openstack/nova master: [placement] Require at least one allocation when PUT https://review.openstack.org/490195 | |
| 20:26:09 | mriedem | ha yup, 1962 | |
| 20:26:19 | cdent | what’s your limit? | |
| 20:26:29 | mriedem | well i got the reference | |
| 20:26:33 | mriedem | so limit=01 | |
| 20:26:34 | mriedem | *-1 | |
| 20:26:57 | mriedem | the beach boys: playing at a state fair near you! | |
| 20:28:01 | cdent | is the doubling stack still in flux or is some of it ready for review? | |
| 20:28:13 | mriedem | i'm doing it | |
| 20:28:52 | cdent | is that yes or no? | |
| 20:29:00 | mriedem | no | |
| 20:29:13 | dansmith | not in flux or not reviewable? | |
| 20:29:20 | dansmith | I think it's worth a review in its current state, personally | |
| 20:29:30 | mriedem | i'm about to push, | |
| 20:29:36 | mriedem | but should i rebase https://review.openstack.org/#/c/488510/ on top? i suppose i should | |
| 20:31:16 | mriedem | here it comes | |
| 20:31:27 | openstackgerrit | Matt Riedemann proposed openstack/nova master: remove provider allocs in confirm/revert resize https://review.openstack.org/488510 | |
| 20:31:27 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Sum allocations in the scheduler when resizing to the same host https://review.openstack.org/490085 | |
| 20:31:28 | openstackgerrit | Matt Riedemann proposed openstack/nova master: Test resize to same host with placement api https://review.openstack.org/489973 | |
| 20:33:47 | sdague | mriedem: has anyone attempted to figure out the toc on the sidebar thing? | |
| 20:34:05 | openstackgerrit | Chris Dent proposed openstack/nova master: [placement] Avoid error log on 405 response https://review.openstack.org/490021 | |
| 20:34:09 | sdague | because otherwise I'll look at that tomorrow, because the current toc is driving me nuts | |
| 20:34:33 | sdague | also, this is our top current 404 - https://review.openstack.org/#/c/490181/ | |
| 20:34:47 | sdague | 56 fails on that in the last 2 days | |
| 20:35:00 | mriedem | sdague: nope | |
| 20:36:28 | mriedem | sdague: you should rebase on top of https://review.openstack.org/#/c/489641/ | |
| 20:36:43 | sdague | mriedem: sure | |
| 20:37:00 | openstackgerrit | Sean Dague proposed openstack/nova master: add top 404 redirect https://review.openstack.org/490181 | |
| 20:39:05 | mriedem | you have to re-sort now too | |
| 20:39:20 | openstackgerrit | Sean Dague proposed openstack/nova master: sort redirectmatch lines https://review.openstack.org/490182 | |
| 20:39:29 | sdague | yep, done | |
| 20:39:42 | mriedem | hot | |
| 20:39:43 | mriedem | dog | |
| 20:40:55 | mriedem | all approved | |
| 20:41:26 | sdague | \o/ | |
| 20:46:10 | openstackgerrit | Sean Dague proposed openstack/nova master: fix rpc broken rst comment https://review.openstack.org/490197 | |
| 20:46:12 | sdague | mriedem: one more for you | |