| *** lakshmiS has joined #openstack-searchlight | 00:43 | |
| *** itisha has quit IRC | 01:03 | |
| *** lakshmiS_ has joined #openstack-searchlight | 01:17 | |
| *** lakshmiS has quit IRC | 01:21 | |
| *** lakshmiS_ has quit IRC | 01:52 | |
| *** lakshmiS has joined #openstack-searchlight | 04:21 | |
| *** lakshmiS has quit IRC | 04:55 | |
| *** lakshmiS has joined #openstack-searchlight | 06:32 | |
| *** lakshmiS has quit IRC | 06:37 | |
| *** ekarlso has quit IRC | 06:58 | |
| *** lakshmiS has joined #openstack-searchlight | 07:24 | |
| *** ekarlso has joined #openstack-searchlight | 07:25 | |
| *** lakshmiS has quit IRC | 11:51 | |
| *** lakshmiS has joined #openstack-searchlight | 12:30 | |
| *** ekarlso has quit IRC | 12:38 | |
| *** lakshmiS_ has joined #openstack-searchlight | 12:53 | |
| *** ekarlso has joined #openstack-searchlight | 12:56 | |
| *** lakshmiS has quit IRC | 12:57 | |
| *** lakshmiS has joined #openstack-searchlight | 13:00 | |
| *** lakshmiS_ has quit IRC | 13:03 | |
| *** lakshmiS has quit IRC | 13:16 | |
| *** lakshmiS has joined #openstack-searchlight | 14:03 | |
| *** sigmavirus24_awa is now known as sigmavirus24 | 14:10 | |
| TravT | hi lakshmiS | 14:55 |
|---|---|---|
| *** itisha has joined #openstack-searchlight | 14:55 | |
| lakshmiS | hi travis | 14:56 |
| TravT | nikhil: sjmc7: sigmavirus24 david-ly_ rosmaita did you see Theirry's response. Sounds like we need to go to the with-milestone model. | 14:56 |
| sjmc7 | yes. it made sense | 14:56 |
| lakshmiS | i guess we dont have other choice | 14:57 |
| TravT | well, the other choice is independent | 14:58 |
| TravT | http://governance.openstack.org/reference/tags/release_independent.html | 14:58 |
| nikhil | yeah | 14:59 |
| nikhil | I think people in general dislike independent model | 14:59 |
| nikhil | but it may not be the case for us | 14:59 |
| sigmavirus24 | I see | 15:00 |
| sigmavirus24 | Yeah I think that may need to be more explicitly stated in their documentation | 15:00 |
| TravT | most the projects look to be infrastructure or plugins | 15:00 |
| nikhil | I would vote for milestone if we can do that easily but that's not super critical from my POV | 15:00 |
| sigmavirus24 | TravT: correct | 15:00 |
| TravT | kinda suprised to see magnum and solum on there. | 15:00 |
| sigmavirus24 | Yeah, ttx's response makes sense to me. | 15:00 |
| sigmavirus24 | TravT: on independent? | 15:00 |
| sigmavirus24 | Well they were both started by Adrian Otto... who is an interesting person | 15:01 |
| TravT | yeah, on independent | 15:01 |
| sjmc7 | indepenent is "do it when we want"? | 15:01 |
| sjmc7 | looks like there're a lot of less mature projects on there | 15:01 |
| TravT | this page says the following about independent | 15:01 |
| TravT | http://docs.openstack.org/project-team-guide/release-management.html | 15:01 |
| TravT | Deliverables that do not benefit from a coordinated release or from stable branches may opt to follow a completely independent release model. | 15:02 |
| TravT | Versions are tagged from the master branch without any specific constraint, although the usage of a post-version numbering scheme based on semantic versioning is strongly recommended. | 15:02 |
| nikhil | TravT: are you talking about the email sub" Liberty release planning help for Searchlight" ? | 15:04 |
| nikhil | I seem to be a bit .. out of sync | 15:04 |
| TravT | we can go with milestones, but my discontent with that model is that since a lot of our project is plugin based, it isn't great to have to go 6 months before releasing new plugins. | 15:04 |
| sjmc7 | maybe independent is better for now then | 15:04 |
| TravT | nikhil: yeah, we had a big discussion on here yesterday and you weren't online. | 15:04 |
| TravT | [openstack-dev] [searchlight] Liberty release finalization | 15:04 |
| TravT | nikhil: ^ that's the email. | 15:04 |
| nikhil | ah, the power was out, oops | 15:04 |
| nikhil | TravT: Thanks! | 15:05 |
| nikhil | I was offline w/ electric for ~3 hours and missed it | 15:05 |
| TravT | ahh, ok | 15:05 |
| *** david-ly_ is now known as david-lyle | 15:06 | |
| TravT | sjmc7 to play devil's advocate to my statement above, the release milestone is a little more conducive early in the project for us to make adjustments across plugins. | 15:06 |
| *** lakshmiS has quit IRC | 15:06 | |
| *** lakshmiS has joined #openstack-searchlight | 15:06 | |
| *** lakshmiS_ has joined #openstack-searchlight | 15:10 | |
| TravT | david-lyle: here's a question for you. which release model will be more acceptable to horizon consuming searchlight? | 15:11 |
| TravT | david-lyle: we don't have a client. we are doing direct calls to searchlight. if it makes sense to just be on the coordinated release with milestones and horizon can consume whatever is latest merged on searchlight, then maybe just doing the milestone release process is easiest to stay in sync with most of the main OpenStack projects. | 15:14 |
| *** lakshmiS has quit IRC | 15:14 | |
| * TravT sees that ttx is talking to david-lyle in the relmgr office | 15:14 | |
| david-lyle | TravT: Horizon should be fine | 15:14 |
| david-lyle | we just need to check the API version | 15:15 |
| david-lyle | and only make calls where the API version is high enough | 15:15 |
| david-lyle | it's actually easier than the client IMO | 15:15 |
| TravT | yeah, i think this means we do need to increment an api micro version anytime we expose a new feature, etc. | 15:17 |
| TravT | initially, it really won't be much of a problem. | 15:17 |
| david-lyle | I think a micro version is best | 15:18 |
| TravT | i'll open a BP on that. | 15:19 |
| TravT | sigmavirus24: lakshmiS_: sjmc7: nikhil: | 15:19 |
| TravT | rosmaita: | 15:19 |
| nikhil | TravT: IIRC, ironic stil does independent release model | 15:19 |
| nikhil | (I read through the emails and ref of the docs) | 15:19 |
| nikhil | one question was -- can we change the rel model later? | 15:20 |
| TravT | i'm sure we can. | 15:20 |
| nikhil | say if we go with independent for now and move to intermediary , would that work? | 15:20 |
| rosmaita | TravT: give me a few minutes to read back | 15:20 |
| nikhil | it will help us in cases if/when we decide to have a client and support diff projects | 15:20 |
| TravT | i was almost thinking the opposite... start with milestone and switch to independent if needed. | 15:21 |
| nikhil | huh :) | 15:21 |
| nikhil | do you mean intermediary? | 15:21 |
| nikhil | as that sorta indicates more maturity | 15:21 |
| nikhil | may be in a cycle or two | 15:21 |
| nikhil | well, my thought process is | 15:22 |
| TravT | no, milestone... ttx was saying intermeidiary is for stable projects | 15:22 |
| TravT | http://docs.openstack.org/project-team-guide/release-management.html | 15:22 |
| TravT | - Common cycle with intermediary releases -> his is especially suitable to more stable projects which add a limited set of new features and don’t plan to go through large architectural changes. Getting the latest and greatest out as often as possible, while ensuring stability and upgradeability. | 15:23 |
| nikhil | TravT: sorry, I meant to say-- if we go with milestone now then switch to intermediary later | 15:23 |
| TravT | ah ok. | 15:23 |
| TravT | yeah, that might make sense | 15:23 |
| TravT | the purpose described for milestone does sound more like where we are at... | 15:23 |
| nikhil | the one drawback with milestone is -- we need to follow other projects a lot | 15:23 |
| nikhil | I see | 15:23 |
| TravT | although it is very odd to me that they describe milestone as being for young projects, when nova is on it. | 15:23 |
| nikhil | milestone requires a LOT of co-ordination | 15:24 |
| nikhil | and it may create a deps hell once we involve prj like docs, translations, may be even nova, glance once the updates to those start affecting searchlight in diff ways | 15:25 |
| david-lyle | I think the the intermediary provides a lot more flexibility | 15:25 |
| david-lyle | which is good when you're attempting to rapidly add support | 15:26 |
| david-lyle | and harden | 15:26 |
| TravT | david-lyle: i thought so too, but in the ML, ttx seemed to discourage us from that. | 15:26 |
| nikhil | I think co-ordination is sorta inherently needed by nova so that model befits them. also, it's a old , tried and tested and possibly they don't want to port to something new and confuse people with more change in the process -- it's tricky as hell as of now | 15:26 |
| david-lyle | ah, haven't read the ML this morning yet | 15:26 |
| david-lyle | I think relmgmt likes milestones | 15:26 |
| nikhil | david-lyle: they want the release numbers to not presume alpha, beta etc as a part of the releases | 15:27 |
| TravT | yeah, check out the [openstack-dev] [searchlight] Liberty release finalization | 15:27 |
| nikhil | for intermediary | 15:27 |
| * david-lyle peeking into the abyss | 15:27 | |
| nikhil | so, intermediary makes sense for stable projects | 15:27 |
| nikhil | TravT: independent is quite flexible... but it's quite infamous whenever brought up at the CPL meetings | 15:28 |
| nikhil | and as Ironic implements that model well, we can see and learn from them. | 15:29 |
| TravT | nikhil: yeah, the release independent sounds great on the surface, but also kind of not so open-stacky | 15:29 |
| nikhil | One good thing with independent is that it will help when releasing only client like changes | 15:29 |
| nikhil | TravT: heh, true not so openstacky is prolly the most fitting word | 15:30 |
| nikhil | phrase* | 15:30 |
| TravT | the thing about us being a lot of plugins is what appealed to me for intermediary / independent. But, here's what i'm thinking right now. | 15:31 |
| TravT | we are facing some uncertainty about what is best, so perhaps it is best to start with the most standard approach (milestones). they seem to be intended to allow for reworking without the expectation of backwards compatibility during the release. We can switch to independent or intermediary as soon as / if we find milestones no longer fit our needs. | 15:33 |
| TravT | i know of at least one deployer that pulls master bits as needed anyway. | 15:35 |
| TravT | nikhil: it seemed like you were leaning towards independent, though? | 15:36 |
| nikhil | TravT: not really inclining towards :) I was trying to mark the tradeoffs and see what fits us best just to ensure we haven't missing any critical point | 15:37 |
| nikhil | TravT: if we are okay with milestone considering all the constraints, nothing like it. It's a bit heavy weight for sure but if we all are on the same page on how to handle it --- I would support it for sure! | 15:38 |
| TravT | in theory, milestone releases give you opportunity to get code in within a release and iterate on it. | 15:39 |
| nikhil | it has it's own value in terms of documenting deliverables and also puts a good face for us to the wider community | 15:39 |
| TravT | although this doesn't seem to be how many projects treat it. | 15:39 |
| nikhil | yes, practical impl vary as there are some loopholes | 15:39 |
| nikhil | and for cases when there are some issues with the API related changes and major refactor, it comes in the way (somewhat). | 15:40 |
| TravT | Well, i think this is a decision that can be changed later and we want to get a liberty release out. | 15:43 |
| TravT | so, i think i'll put out an rc1 tag (assuming milestone release process). | 15:43 |
| nikhil | TravT: works for me! | 15:44 |
| TravT | any veto's sjmc7: david-lyle: lakshmiS_: rosmaita | 15:44 |
| lakshmiS_ | nope | 15:45 |
| david-lyle | either release model is fine with me | 15:45 |
| TravT | sjmc7: lakshmiS_: FYI i did find one bug with facet options yesterday. it is minor. | 15:46 |
| TravT | https://bugs.launchpad.net/searchlight/+bug/1503080 | 15:46 |
| openstack | Launchpad bug 1503080 in OpenStack Search (Searchlight) "Facet doc count reports more docs than actual" [Undecided,New] | 15:46 |
| TravT | nothing to hold release over, but would be nice to fix since I'm now displaying that in the horizon plugin | 15:46 |
| nikhil | itisha: you have update on your config related patch? | 15:47 |
| lakshmiS_ | can it be backported after release? | 15:47 |
| sjmc7 | ah, because they're nested? | 15:50 |
| TravT | i don't know... after i added that I jumped into adding a settings button to change things like limit, etc. | 15:51 |
| sjmc7 | ok. yeah, that's an interesting one | 15:51 |
| TravT | lakshmiS_: we can backport that if needed. | 15:57 |
| TravT | since it is a fix. | 15:57 |
| lakshmiS_ | have to do regression test again with any more changes | 15:58 |
| TravT | i think we should just push the rc1 tag | 16:01 |
| TravT | and if there is a need to backport, we will | 16:01 |
| lakshmiS_ | yeah agree | 16:02 |
| TravT | Ok, tag applied! Will now follow up with ttx for next steps. | 16:16 |
| TravT | https://github.com/openstack/searchlight/releases | 16:17 |
| lakshmiS_ | great! | 16:23 |
| *** lakshmiS_ has quit IRC | 16:44 | |
| *** david-ly_ has joined #openstack-searchlight | 17:21 | |
| *** david-lyle has quit IRC | 17:21 | |
| *** david-ly_ is now known as david-lyle | 17:22 | |
| *** akanksha_ has joined #openstack-searchlight | 20:14 | |
| *** sigmavirus24 has quit IRC | 20:43 | |
| *** pkarikh has quit IRC | 20:43 | |
| *** sigmavirus24 has joined #openstack-searchlight | 20:46 | |
| *** pkarikh has joined #openstack-searchlight | 20:46 | |
| openstackgerrit | Travis Tripp proposed openstack/searchlight: Update setup.cfg to version 0.1.0 https://review.openstack.org/231719 | 21:23 |
| TravT | ^ we just got a stable/liberty branch | 21:24 |
| sjmc7 | woo! | 21:24 |
| sjmc7 | good job schmoozing! | 21:24 |
| TravT | but found that we missed getting setup.cfg to the correct version | 21:24 |
| TravT | well crap | 21:25 |
| TravT | i did it on the wrong branch | 21:25 |
| sjmc7 | *sad trombone* | 21:25 |
| openstackgerrit | Travis Tripp proposed openstack/searchlight: Update setup.cfg to version 0.1.0 https://review.openstack.org/231724 | 21:33 |
| TravT | sjmc7: can you see if you have +2 rights on https://review.openstack.org/#/c/231726/1 | 21:51 |
| TravT | dhellmann said we need to get that pushed through right away. | 21:51 |
| david-lyle | TravT: after the fact +2, but now you have it | 21:55 |
| TravT | david-lyle: thanks! | 21:55 |
| sjmc7 | i don't | 22:00 |
| sjmc7 | doug was like quick draw mcgraw | 22:00 |
| TravT | okay there must be ACL's file we need to update. I think we need to give the release management team some permissions to | 22:01 |
| TravT | too | 22:01 |
| david-lyle | oh, I'm already on stable | 22:04 |
| david-lyle | that's why | 22:04 |
| david-lyle | not related to searchlight | 22:04 |
| david-lyle | there should be a stable-maint ACL | 22:04 |
| *** itisha has quit IRC | 22:09 | |
| openstackgerrit | Travis Tripp proposed openstack/searchlight: Update setup.cfg 0.2.0 https://review.openstack.org/231739 | 22:17 |
| TravT | ^ mitaka version | 22:17 |
| TravT | see commit comments, can discuss more if needed | 22:17 |
| openstackgerrit | Merged openstack/searchlight: Glance exceptions should be removed https://review.openstack.org/230086 | 22:34 |
| *** itisha has joined #openstack-searchlight | 22:44 | |
| *** sigmavirus24 is now known as sigmavirus24_awa | 23:16 | |
| TravT | FYI, I've added searchlight release notes to the liberty release notes wiki. Please check to see if I missed anything. https://wiki.openstack.org/wiki/ReleaseNotes/Liberty#OpenStack_Search_.28Searchlight.29 | 23:39 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!