14:00:30 #startmeeting glance 14:00:31 Meeting started Thu Dec 14 14:00:30 2017 UTC and is due to finish in 60 minutes. The chair is rosmaita. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:00:32 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 14:00:35 The meeting name has been set to 'glance' 14:00:41 #topic roll call 14:00:42 o/ 14:00:48 o/ 14:01:36 o/ 14:01:40 o/ 14:02:38 o/ 14:03:14 ok, looks like we have a quorum 14:03:20 good morning/afternoon/evening ,everyone 14:03:35 as always, here's the agenda: 14:03:36 hey 14:03:41 #link https://etherpad.openstack.org/p/glance-team-meeting-agenda 14:03:41 good morning rosmaita 14:03:53 and it is a very light agenda today 14:04:07 so plenty of time for open discussion 14:04:14 #topic updates 14:04:35 ok, so i was supposed to do a few things this week, namely: 14:04:47 spec about policy changes 14:04:56 no-op community goal proposal 14:05:18 unfortunately, most of my time was consumed with the python-glanceclient gate situation 14:05:36 so i need to check what the deadline is for community goal proposals 14:05:41 (i hope it is not soon) 14:06:01 and i've got the policy changes partially written up, but not pushed to gerrit yet 14:06:17 so, the python-glanceclient gate situation 14:06:39 it's been broken for a while as a casualty of the zuul3 migration 14:06:56 some test jobs were auto-migrated in such a way that they no longer worked 14:07:05 and that's what happened to us 14:07:37 the nice thing about zuul3 is that we can define our gates in our own repo 14:08:20 so there's a patch up moving the jobs to the python-glanceclient repo and rewriting them to use zuul3 and work 14:08:41 that's mostly done, though it depends on a patch in the devstack repo that hasn't merged yet 14:08:45 (last time i looked) 14:08:48 unfortunately 14:09:00 (how many times have i used that word?) 14:09:12 which is still waiting for the Monty's patch that is being bikeshed under devstack 14:09:27 in between when the functional test jobs were passing previously and now, some stuff has broken 14:09:48 like? 14:09:49 i have a patch up (that i think merged yesterday) to fix one problem 14:10:00 i'm almost tempted to say that we grab that Monty's patch, to our own repo and consume it locally rather than wait much longer 14:10:24 but now the funcitonal jobs don't run in devstack when the tls-proxy is running 14:10:44 i.e., when you're using ssl, the tests break 14:11:11 got it 14:11:14 i filed a bug, i'll post the number in a minute 14:11:28 it's a cert problem, but i haven't been able to isolate exactly where it's happening yet 14:11:44 except really early, the client can't connect to keystone 14:13:22 sorry, was looking in the glance bugs by mistake 14:13:24 https://bugs.launchpad.net/python-glanceclient/+bug/1738033 14:13:30 Launchpad bug 1738033 in Glance Client "functional tests broken in ssl environment" [Critical,Triaged] - Assigned to Brian Rosmaita (brian-rosmaita) 14:14:02 it's assigned to me, but i'll be happy to hand it over if someone else wants it 14:14:22 also, i will be happy to get advice from anyone familiar with this kind of problem 14:14:48 jokke_ i think you have a good suggestion 14:14:54 about monty's patch 14:15:06 I am not familiar, but I will have a look tomorrow 14:15:19 apparently the patch that we're depending on breaks some other project's job 14:15:35 so if we move it to our repo, only we will be using it 14:15:56 we can always revert to using the job def in the devstack repo once that's sorted out 14:16:00 I'm sure that someone will start crying that we're doing code duplication as soon as it merges, but looking the comment's I think santa comes first than we have tests running if we wait that 14:16:24 exactly 14:16:38 ok, i will do that right after the meeting 14:17:30 so to summarize, the situation is that right now we are running with ZERO functional test gates for glanceclient 14:18:02 the patches we have now will restore the gates, but only run hte functional tests without ssl 14:18:14 which is better than nothing, but also far from ideal 14:18:37 what i'd like to do is: 14:18:48 (1) get the functional test gates running again 14:18:59 (2) then we can merge the backlog of glanceclient changes 14:19:14 (3) fix it so the functional gate works under ssl 14:19:29 and not cut a release of the python-glanceclient until we have (3) done 14:19:57 any comments? 14:20:16 no, sounds good 14:20:31 #action rosmaita edit current functional gate patch to include monty's new devstack-base job 14:20:38 please let me know if I need to look during my day time 14:20:40 yeah, I'd be very uncomfortable releasing without ssl being tested 14:20:55 once i have that up, it's all on us, which is the nice thing about zuul3 14:21:07 also I think finding possible solution to fix that ssl issue gets easier if we have gate that is able to run those tests 14:21:25 abhishekk i'll shoot you an email 14:21:36 rosmaita: sure 14:21:45 i should have everything ready to go so that you can look at it in your morning 14:22:01 also, i'll send you the little info i learned while looking at the ssl problem 14:22:16 sounds good, that will be helpful 14:22:44 but it's like 8 pm your time, so please enjoy your evening after the meeting 14:22:48 :) 14:22:54 ok, that's all from me 14:22:57 :) 14:23:08 gimme a sec to find the patch i'm talking about so everyone knows 14:24:17 got it 14:24:22 #link https://review.openstack.org/#/c/525045/ 14:24:42 ok, so i'll revise and rebase that patch, it should eliminate both dependencies 14:25:14 one of the dependencies is monty's patch to the playbooks in the devstack repo 14:25:37 (in case you're interested in looking at it) 14:25:48 ok, that really is all from me 14:25:58 actually, one more thing 14:26:29 i will be travelling this weekend and only working limited hours monday-wednesday next week 14:26:45 k 14:26:53 noted 14:26:57 One thing from me ref the client release 14:26:59 thanks 14:27:01 ok 14:27:02 ok 14:27:15 jokke_ go ahead 14:27:24 please do make sure that if you have anything in flight for the client it get reviewed and ready to go 14:27:38 ok 14:27:44 we want to merge the things we include and release asap after the gate is rectified 14:28:28 there's lots queued up waiting to be released that has been fixed this cycle 14:28:38 so about that ... we had talked about releasing out of stable/pike ? 14:28:56 i have not checked what's happening with the test jobs in stable/pike 14:29:06 i am afraid it may not be good 14:29:20 may have to backport this fix to stable/pike 14:29:22 me neither, but I assume we have same shit waiting there 14:29:37 what jokke_ said 14:30:01 it's not like stable branches wouldn't use zuul 3 14:30:48 right, and the plan is that once we define jobs in the repo, the jobs will go with the stable branch when they're cut, and that will allow us to change jobs in master without worrying about stable branches no longer passing 14:31:48 ok, any other updates? 14:31:54 that is good point 14:32:12 I can share an update 14:32:18 it's very dangerous as well ... so we need to have sharp eye on the test changes to make sure we don't cause any regressions 14:32:47 jokke_ ++ 14:32:54 McClymontS go ahead 14:33:14 The multihash spec is ready to be looked at for the most part 14:33:29 you mean the implementation! 14:33:45 unfortunately, I hit a snag late in the day yesterday and had to go so its not fully on gerrit yet but hoping to get the glance store element is expected to be on there soon 14:34:03 that's great 14:34:06 yep! the code is mostly there, we will need to discuss a few things in the open forum here but I think that will have to be next week now 14:34:20 why don't you send an email to the ML when you have the patches up 14:34:37 Specifically log levels, if any additional logging is wanted and how we exactly want to go about defaulting to sha if checksum only exists for many older images 14:34:39 and then we can start looking 14:34:44 will do, just wanted everyone to know its very close 14:35:19 ok, put yourself on the agenda for next week and we can discuss 14:35:27 glad to hear it's close! that's good work 14:35:27 sounds good 14:35:41 yes thanks, excited to get a feature in 14:36:01 with all the glance related OSSNs out there, it's nice to have a new security feature 14:37:27 ok, i guess that's all for updates 14:37:35 #topic open discussion 14:37:44 Yesterday I have submitted patch [1] for Inject image metadata feature Implementation (https://blueprints.launchpad.net/glance/+spec/inject-automatic-metadata) [1]: https://review.openstack.org/#/c/527635/1 abhishekk reviewed the patch and given suggestion (Split out patch in 3 different patches) to avoid zuul tempest failing 14:37:53 Submitted 3 patches 14:38:14 [1]: https://review.openstack.org/#/c/527906/ (create glance-image-import.conf file all changes regarding new file will go in this patch, also you need to add sample file under etc directory as glance-image-import.conf and not as glance-image-import.conf.sample) 14:38:22 [2]: https://review.openstack.org/#/c/527907/ (devstack to copy glance-image-import.conf file to /etc/glance directory) 14:38:30 [3]: https://review.openstack.org/#/c/527635/2 (Code to inject metadata properties) But this patch requires to merge above two patches. 14:38:37 that was my opinion though, jokke_, rosmaita please share your views as well 14:38:38 So I would like to see more opinions and reviews on above patches. 14:39:13 rosmaita: I will need your help to review [2] I mean in devstack we are making changes in lib/glance 14:39:55 saw the comments given by abhishekk on [1] and [2] mentioned above i will fix as early as possible. thanks abhishekk!! 14:40:20 bhagyashris: no issues 14:41:23 quick question about [2] 14:41:50 actually, i will look at the patches first 14:42:18 i have a patch up removing copying of conf files from devstack, instead use iniset to create the conf on the fly 14:42:39 which is what we do for most of the conf files currently in devstack 14:43:04 #link https://review.openstack.org/#/c/527263/ 14:43:20 anyway, maybe we can do that for this as well 14:43:37 ok 14:43:41 ok, nice work bhagyashris, i look forward to looking over your patches! 14:44:07 anything else? 14:44:19 so should i make changes in the devstack patch as per https://review.openstack.org/#/c/527263/ this? 14:44:38 if that makes sense for what you're doing 14:44:54 ok 14:45:07 take a look and see what you think 14:45:32 yes sure 14:45:46 thank you :) 14:46:08 So for which milestone we are targeting policy related changes?, q-3 or r-*? I guess of rocky though 14:46:20 abhishekk will have to be rocky 14:46:42 rosmaita: got it 14:46:52 bhagyashris: I'll try to leave my comments in there today or tomorrow morn 14:47:13 bhagyashris: I have couple principal concerns about the implementation 14:47:42 jokke_: ok thank you :) 14:48:11 anything else for open discussion? 14:48:37 I'd like to ask about peoples holiday plans 14:48:53 so we're just a week away from Xmas 14:49:04 I will be having holiday on 25 and 1 14:49:33 same me too 14:49:38 I will personally be around at Xmas time, but likely be travelling from start of Jan 14:49:43 I'm mostly out the 25th to 3rd, but will be in and out if anyone needs anything from me. 14:49:50 i will be mostly away from dec 25 to jan 4 14:50:14 like smcginnis i will be occasionally online doing stuff 14:50:49 ok ... so if something critical happens myself and abhishekk are the main points of contacts between Xmas and new year then 14:51:07 yes, possibly nikhil as well 14:51:09 good to know that we have at least 2 cores more or less available :D 14:51:10 yup 14:51:27 I'm in and out from 22-5 14:51:55 ok, if there's nothing else, we can end early today 14:52:09 Thanks all! 14:52:13 Mostly we will not have meeting during xmas week then 14:52:15 thank you 14:52:17 thanks everyone! 14:52:32 thank you 14:52:36 abhishekk good point 14:52:47 should we meet on dec 28 ? 14:52:49 yeah, no meeting at 28th 14:52:58 ack 14:53:02 ok, we can cancel it now 14:53:07 no meeting on dec 28 14:53:08 sounds good 14:53:18 but, will look forward to seeing you all next week! 14:53:25 #endmeeting