| Posted | Nick | Remark | |
|---|---|---|---|
| #openstack-nova - 2017-10-24 | |||
| 13:41:40 | mriedem | oh | |
| 13:42:16 | belmoreira | all the services that we have are global | |
| 13:42:53 | dansmith | belmoreira: I would check with keystone people to see if there's a way you can avoid having to deploy a new keystone everywhere | |
| 13:42:57 | belmoreira | dansmith: good point... maybe I can do something with endpoint filtering | |
| 13:43:09 | dansmith | also, you could do some DNS hackery to make the same url be "different" for each cell | |
| 13:43:18 | dansmith | if keystone can't help | |
| 13:43:58 | mriedem | could you do a placement endpoint per region and make a region per cell? | |
| 13:44:47 | mriedem | but just for the placement service | |
| 13:45:11 | belmoreira | mriedem: I see your point. That may work. thanks | |
| 13:45:21 | mriedem | so everything talks to cinder/neutron/glance on RegionOne, | |
| 13:45:30 | mriedem | and nova-compute in cell1 talks to PlacementRegionCell1 or something | |
| 13:46:10 | dansmith | yeah | |
| 13:46:18 | belmoreira | mriedem: makes sense. That may be the easiest solution. thanks | |
| 13:46:26 | dansmith | belmoreira: basically we're saying, we hope you don't have to deploy more keystone ;P | |
| 13:46:40 | belmoreira | :) | |
| 13:46:41 | sdague | mriedem: yeh, let me figure out why the tests aren't passing | |
| 13:46:54 | mriedem | sdague: one was cellsv1 random failure as usual, | |
| 13:47:01 | mriedem | other was zuulv3 reno job | |
| 13:47:04 | mriedem | but logs are gone | |
| 13:47:18 | mriedem | belmoreira: and i hope you're taking notes so you can post a blog on all of this when it's done | |
| 13:47:23 | mriedem | that would be awesome | |
| 13:47:58 | sdague | mriedem: ok, I just rechecked on pike, once that is in, I'll approve ocata, just don't want them to go in in the wrong order | |
| 13:48:14 | mriedem | yup, thanks | |
| 13:48:16 | belmoreira | yeah. are you aware of any cellV1 to V2 migration already? | |
| 13:48:33 | mriedem | belmoreira: i haven't heard of any | |
| 13:48:47 | mriedem | spamaps at godaddy said they were talking about just standing up a new pike cells v2 deployment | |
| 13:48:48 | mriedem | rather than migrate | |
| 13:49:19 | belmoreira | lucky ones | |
| 13:49:20 | dansmith | not sure that's because migration is hard so much as a chance to start fresh? | |
| 13:49:36 | mriedem | he mentioned while talking about their normal upgrade pain of carrying patches forward | |
| 13:49:43 | mriedem | so it could just be a chance to start fresh yeah | |
| 13:49:48 | sdague | stephenfin: https://review.openstack.org/#/c/495541 would land another 3 privsep patches | |
| 13:50:16 | mriedem | i haven't heard from nectar | |
| 13:50:18 | mriedem | but never really do | |
| 13:50:24 | mriedem | mgagne: you're cellsv1 aren't you? | |
| 13:51:29 | belmoreira | also, in newton there's the placement_database group in the configuration. And it's advised to use it. I don't see any reference of it now in master | |
| 13:51:32 | stephenfin | sdague: done and done | |
| 13:51:46 | mriedem | belmoreira: that's a mistake, don't use that | |
| 13:52:10 | mriedem | i thought we cleaned that all out actually | |
| 13:52:53 | mriedem | ah we didn't revert that from newton https://github.com/openstack/nova/commit/1b5f9f8203c90fe447d33c89f238104026052d1e | |
| 13:52:59 | stephenfin | sdague: Regarding https://review.openstack.org/#/c/500354/, any reason you think we were using 'discard_warnings' before? | |
| 13:53:18 | stephenfin | i.e. what's the chances we'd regress on something by dropping it | |
| 13:53:42 | mriedem | newton isn't eol yet, i could still revert that before eol... | |
| 13:54:10 | mriedem | yeah we removed in ocata https://review.openstack.org/#/c/442762/ | |
| 13:54:16 | belmoreira | mriedem: ok | |
| 13:55:38 | mriedem | this was the revert https://review.openstack.org/#/c/360638/ | |
| 13:55:58 | mriedem | ah yeah so the revert removed the reno and code, but missed the config options, | |
| 13:56:04 | mriedem | which is what my change removed | |
| 13:56:12 | mriedem | so we should consider backporting that too newton also before the eol | |
| 13:57:42 | stephenfin | dansmith: Got a sec to discuss https://review.openstack.org/#/c/509107/ ? | |
| 13:57:47 | belmoreira | thanks for all the suggestions | |
| 13:58:06 | belmoreira | after moving to Ocata with this configuration (means that I still keep cellsV1 schedule behaviour) then for Pike the plan is to remove all these placements and move the existing cellsV1 to cellsV2 (point everything to the global nova_api and do the mappings) | |
| 13:58:15 | dansmith | stephenfin: yup | |
| 13:58:20 | stephenfin | I was the one that actually told ralonsoh to drop the versioning code in a dependent patch because I didn't think it was necessary | |
| 13:58:37 | belmoreira | is there any know blocker that I'm missing or this should work then? | |
| 13:58:39 | dansmith | belmoreira: cool | |
| 13:58:39 | stephenfin | as we don't store that in a database nor do we send it between services | |
| 13:58:41 | mriedem | tonyb: what's the status on newton eol? because i think i've got another worthwhile change for newton before eol | |
| 13:59:10 | mriedem | belmoreira: honestly we probably won't know much about cellsv1 issues here as none of us are running it | |
| 13:59:23 | dansmith | stephenfin: if we register it then we really shouldn't be breaking the rules because if someone were to send us a 1.0 version of that we couldn't even deserialize it | |
| 13:59:43 | dansmith | stephenfin: if we feel it's overly onerous to break the rules, then we might as well not even bump the version because we *can't* honor the old one | |
| 13:59:53 | dansmith | overly onerous to *not* break the rules Imean | |
| 14:01:12 | stephenfin | dansmith: Yeah, that makes sense. I guess we might want to revert part of https://review.openstack.org/#/c/508498/ too, in that case | |
| 14:01:27 | belmoreira | mriedem: dansmith: ok. thanks, I will go through these steps and I will let you know this goes | |
| 14:01:34 | stephenfin | (that's the one where I said not to do this versioning against ralonsoh's better instincts :)) | |
| 14:02:05 | dansmith | stephenfin: eesh | |
| 14:03:25 | dansmith | stephenfin: well, it's done now, so maybe just reset the versions to 1.0, admit fault, and try to be clean going forward | |
| 14:03:40 | dansmith | reverting an unsupportable object version isn't particularly helpful either | |
| 14:03:55 | dansmith | we clearly need a little more oversight on those objects I think | |
| 14:04:52 | dansmith | if we start taking those over the REST API we have to be even more squeaky clean than we are with our RPC APIs, since we don't control all the sending and receiving code together | |
| 14:05:10 | stephenfin | dansmith: For my own clarification, do you need to register objects if you do not want to send them over the wire? | |
| 14:06:03 | dansmith | stephenfin: you need to register them in order to receive them through the automatic deserializer, so it knows where to find an object implementation to use to deserialize a primitive | |
| 14:06:10 | sahid | dansmith: we are goin gto start that at a particular version, right? i'm interested to know what is the "someone" who is going to send a version 1.0? | |
| 14:06:28 | dansmith | stephenfin: if you don't need to do that, you still need to build the object, but not register it | |
| 14:07:30 | dansmith | sahid: I understand that there is no code that should be sending the object at the earlier version with that field set right now. That doesn't mean we just make tons of random changes with versions we can't support until some arbitrary point at which we say things are now stable. That version history will confuse people and provide counterexamples they may copy. | |
| 14:07:38 | stephenfin | dansmith: Hmm, to the best of my knowledge these are used only by nova rn, so we're not sending these over RPC nor the REST API. Therefore deserialization is not yet necessary | |
| 14:08:19 | stephenfin | Yeah, we should either be treating these as versioned objects or not doing so | |
| 14:08:20 | dansmith | stephenfin: yeah, so, if that's the case, we should be not registering them (which won't track their hashes) and leave them at 1.0 instead of recording unsupportable versioned changes | |
| 14:08:30 | dansmith | stephenfin: the problem is, | |
| 14:08:38 | dansmith | stephenfin: this is a library that anything could be picking up and using at any point, | |
| 14:08:54 | dansmith | so if you don't register/version them properly, then nobody can do that until you do | |
| 14:09:09 | openstackgerrit | Matt Riedemann proposed openstack/nova stable/newton: Remove unused placement_database config options https://review.openstack.org/514685 | |
| 14:09:09 | dansmith | so maybe that means until release 2.0 of the library it's not usable across projects, which is okay I guess, | |
| 14:09:11 | mriedem | dansmith: sdague: johnthetubaguy: belmoreira: tonyb: ^ i think we should get that into newton before eol | |
| 14:09:27 | sahid | dansmith: we also have moved like that in os-vif, i'm not sure how that is going to confuse people. | |
| 14:09:31 | dansmith | but it's another weird thing you'd want to communicate to your users | |
| 14:09:57 | sahid | s/also/always | |
| 14:10:39 | dansmith | sahid: you have always had broken versioning? are you really arguing that "consistently broken" is less confusing or better? :) | |
| 14:11:01 | stephenfin | dansmith: So ultimately, it sounds like the easiest thing to do would be to introduce the changes I had ralonsoh drop from PS2 https://review.openstack.org/#/c/508498/2..4/os_vif/objects/subnet.py | |
| 14:11:04 | sahid | dansmith: well it's your point of view | |
| 14:11:23 | dansmith | library interfaces should be stable. library interfaces that exist to provide cross-service communication primitives should be .... *really* stable :) | |
| 14:12:23 | sahid | not sure i do understand what you try to say, we change a field so we bump the version, that seem simple | |
| 14:12:28 | stephenfin | +1 Consider it a point learned :) I'll go add those missing versioning functions and hope no one's deploying os-vif from master, heh | |
| 14:12:53 | dansmith | sahid: because just recording a version does not do anything if you can't support the old one | |
| 14:13:17 | dansmith | sahid: it's not a commit log, it serves an actual purpose, unless you break the rules and then it does nothing but communicates to people something you can't do | |
| 14:13:27 | ralonsoh | stephenfin: should I add a make_compatible in https://review.openstack.org/#/c/508498? | |
| 14:13:39 | sahid | it's just to keep history as a chnagelog is doing | |
| 14:13:41 | ralonsoh | stephenfin: and in https://review.openstack.org/#/c/509107/? | |
| 14:13:48 | dansmith | sahid: it is not | |
| 14:14:38 | sahid | okay okay you win, i'm not going to arguing all the day for something so minor :) | |