| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2020-06-18 | |||
| 14:56:49 | dansmith | personally, I do not understand how relying on a centralized (even if HA'd) load balancer is better than the clients knowing the options and being able to help themselves out of a failure to talk to one | |
| 14:56:51 | mordred | if this is a thing that has to stay around | |
| 14:56:52 | efried | Easy for me to say, I won't be the one doing it. | |
| 14:57:01 | sean-k-mooney | dansmith: ok that more hevy wait then kolla does so i guess that is different in ooo | |
| 14:57:15 | mordred | then I want to just say screw it and support it in the client layer | |
| 14:57:34 | efried | dansmith: ftr, api_servers isn't really a load-balancer. I think it's a deafdumbandblind round robin. | |
| 14:57:36 | mordred | it's way too fundamental of a piece of config to stich in in the way it's being done | |
| 14:57:44 | mordred | a very dumb one | |
| 14:57:46 | mordred | it doesn't do retries | |
| 14:57:53 | efried | or failovers, or anything. | |
| 14:57:55 | mordred | yah | |
| 14:57:59 | efried | You simply get the next one on each subsequent api req | |
| 14:58:14 | mordred | it's probalby _less_ HA | |
| 14:58:17 | efried | heh | |
| 14:58:22 | efried | indeed | |
| 14:58:37 | mordred | maybe it's time to one-more-time go to the mailing list | |
| 14:58:40 | dansmith | I understand, but it still gets you 66% success instead of zero if 1/3 is down, and it could be made to do the right thing | |
| 14:58:44 | mordred | maybe life has changed for peopel since queens | |
| 14:58:56 | dansmith | mordred: this was not queens when we had this discussion, fwiw | |
| 14:58:59 | sean-k-mooney | mordred: for the better? | |
| 14:59:09 | mordred | dansmith: no? that's the release mentioned in the nova config docs | |
| 14:59:14 | dansmith | mordred: anyway, don't use my name in the email, I'm not prepared to argue for it anymore | |
| 14:59:21 | mordred | dansmith: I won't | |
| 14:59:28 | dansmith | mordred: that's when it was deprecated I think, but the concern about removing it was within the last year | |
| 14:59:34 | mordred | nod | |
| 14:59:49 | dansmith | but also, we were told that it was going to go away regardless so those people are assuming it's going or gone I think | |
| 15:00:06 | mordred | oh - well that's something perhaps | |
| 15:00:27 | artom | stephenfin, did the flavor extra spec validation spec ever go anywhere? | |
| 15:00:38 | stephenfin | artom: it landed last cycle | |
| 15:00:47 | stephenfin | API microversion 2.84, I think? | |
| 15:01:43 | artom | stephenfin, 2.86, but yeah, thanks for the pointer | |
| 15:01:56 | artom | So we could use that for the cpu_dedicated_mask, right? | |
| 15:02:08 | efried | Deprecate: https://review.opendev.org/#/c/692227/ (merged) | |
| 15:02:08 | efried | I don't see a patch (at least owned by me) proposing removing it. | |
| 15:02:09 | artom | Context is my latest comment on https://review.opendev.org/#/c/468203/15 | |
| 15:03:19 | efried | Comments in above patch contain ML links mordred | |
| 15:03:45 | efried | and also refers to some people who wanted to keep it. | |
| 15:04:04 | stephenfin | artom: fair comment. In the middle of something at the moment but I'll get a reply to it | |
| 15:04:30 | stephenfin | we can't rely on the validation though, mind you, since it's an API change | |
| 15:04:30 | artom | stephenfin, no rush. Today's my "upstream day", trying to do some reviews where I can :) | |
| 15:04:42 | stephenfin | *in an API microversion | |
| 15:07:55 | mordred | efried: ooh! so it is deprecated! | |
| 15:08:42 | mordred | efried: how many releases do we have to wait to remove a deprecated thing? can we remove that now in victoria? or do we have to wait until w? | |
| 15:09:25 | dansmith | mordred: should probably ask the ptl | |
| 15:10:15 | stephenfin | mordred: it's six months or one release, whatever is greater | |
| 15:10:29 | stephenfin | though that's obviously tempered by impact | |
| 15:10:35 | dansmith | I thought it was two releases for config options | |
| 15:10:49 | dansmith | you can't remove it in the N+1 release, but you can in the N+2, | |
| 15:10:59 | dansmith | else we'd break N->N+1 config compatibility | |
| 15:11:05 | mordred | gotcha. so we can remove in W | |
| 15:11:11 | mordred | but not today | |
| 15:12:16 | stephenfin | you deprecate in N, and remove in N+1 | |
| 15:12:45 | stephenfin | if you removed in N, you'd be breaking N-1 -> N compatibility | |
| 15:13:22 | stephenfin | that gives people the duration of N to get off $deprecatedthing | |
| 15:13:34 | dansmith | isn't N-1->N+1 the same thing I said, but with N->N+2? | |
| 15:14:45 | dansmith | oh, you're saying we deprecated it in U so people in U should know to get off of it so it can be gone in V? I guess technically that's right, but I really thought it was supposed to be two releases | |
| 15:15:03 | stephenfin | yes :) you had me confused there | |
| 15:15:19 | stephenfin | docs say the same https://docs.openstack.org/nova/latest/contributor/process#smooth-upgrades | |
| 15:15:25 | stephenfin | i.e. continue to support and test features for at least one release before they are removed | |
| 15:16:10 | sean-k-mooney | ya so 1 release is the miniume we usuall take 2-3 to remove it | |
| 15:16:16 | stephenfin | definitely not black and white | |
| 15:16:18 | sean-k-mooney | depeneing on how much of a pain it is | |
| 15:16:24 | stephenfin | sean-k-mooney: if we ever remove it :) | |
| 15:16:31 | dansmith | given it could break people's glance setup without warning, | |
| 15:16:34 | dansmith | I'd say it's pretty high impact | |
| 15:16:46 | stephenfin | "will be removed in a future release" <-- my boilerplate | |
| 15:16:52 | sean-k-mooney | we would need a nova-status check at a minium right | |
| 15:17:26 | stephenfin | I wouldn't say without warning, since we do have a big "this thing is deprecated" warning | |
| 15:17:27 | sean-k-mooney | also ooo would have to be modified proably | |
| 15:17:59 | dansmith | sean-k-mooney: for something like this, yeah I'd think so | |
| 15:18:00 | stephenfin | I guess it comes down to how difficult the migration is | |
| 15:18:14 | stephenfin | from $old_way to $new_way | |
| 15:18:30 | stephenfin | I don't know about that so I'll defer to others | |
| 15:25:37 | sean-k-mooney | old way is template a config option. new way is either add the service to a load blancer or deploy a new loadbalncer and add it to that | |
| 15:26:47 | sean-k-mooney | i think the only thing we use this for in ooo wa rabbitmq but i could be wrong about that | |
| 15:27:01 | sean-k-mooney | i think in ooo we go to glance vai haproxy | |
| 15:27:27 | sean-k-mooney | but in general people could have been using this for any of the api endpoint? | |
| 15:47:31 | gibi | nova weekly meeting will start in 12 minutes on #openstack-meeting-3 | |
| 15:49:04 | mordred | sean-k-mooney: well - we wouldnt' be removing this for rabbit | |
| 15:49:13 | mordred | sean-k-mooney: this is only about talkig to openstack api services | |
| 15:49:48 | mordred | it's possible we've not been properly clear about that | |
| 15:50:21 | sean-k-mooney | mordred: why would we remove it for one and not the other | |
| 15:50:34 | sean-k-mooney | its equally bad in both cases | |
| 15:50:48 | sean-k-mooney | or good depending on your view | |
| 15:50:49 | mordred | sean-k-mooney: "use services from the keystone catalog and via the normal api consumption channels" is the driver | |
| 15:51:14 | mordred | rabbit doesn't get configured via the catalog nor consumed using a standard set of connection information | |
| 15:53:38 | sean-k-mooney | mordred: have you tought about a keystone feature to allow the same behavior :P | |
| 15:54:30 | mordred | yes - it was _strongly_ opposed by the keystone team - becaues they did not want to build a bad load balancer | |
| 15:54:32 | sean-k-mooney | but ok if its the openstack services only im not sure if that changes the edge configuration | |
| 15:54:50 | mordred | yeah - and really, it's ONLY glance | |
| 15:54:59 | mordred | that's the only place this option exists | |
| 15:55:27 | mordred | so it's not even that there is a generalized mechanism for talking to api services with a list of endpoits in nova -- it's _just_ for glance | |
| 15:55:50 | mordred | and I'm pretty sure the reason it's special is that way back in the day glance wasn't considered user-facing, so why would you have it in your load balancer | |
| 15:56:40 | sean-k-mooney | well in the edge configurtion you want your edge site to use a local glance api | |
| 15:56:55 | sean-k-mooney | and leavge the glance multistore feature | |
| 15:57:21 | sean-k-mooney | now i dont know if that needs this | |
| 15:57:54 | sean-k-mooney | or can we configure the glance store seperatly and have multiple endpoints in keystone | |
| 15:58:15 | sean-k-mooney | but i feel like we would need the later? | |
| 15:59:07 | sean-k-mooney | you dont want just one endpoint as you would have to do anycast routing tricks if you wanted to hit the local api | |
| 16:10:27 | mordred | sean-k-mooney: well - for those you can totally still use endpoint override | |