00:01:32 <thinrichs> #startmeeting CongressTeamMeeting
00:01:33 <openstack> Meeting started Thu Jul 14 00:01:32 2016 UTC and is due to finish in 60 minutes.  The chair is thinrichs. Information about MeetBot at http://wiki.debian.org/MeetBot.
00:01:34 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
00:01:37 <openstack> The meeting name has been set to 'congressteammeeting'
00:01:50 <aimeeu> hello everyone
00:01:50 <masahito> hi
00:01:53 <ramineni_> hi
00:02:01 <ekcs> hi
00:02:20 <thinrichs> Hi all
00:02:25 <thinrichs> Agenda for today…
00:02:29 <thinrichs> 1. Newton 2
00:02:37 <thinrichs> 2. Barcelona talks
00:02:39 <thinrichs> 3. Status
00:02:42 <thinrichs> Anything else?
00:03:29 <ekcs> gating issue.
00:04:12 <thinrichs> ok
00:04:26 <thinrichs> One other thing.. using Reno for release notes
00:04:30 <thinrichs> 4. Gating
00:04:35 <thinrichs> 5. Reno
00:04:42 <thinrichs> Let's knock a couple of these out quick...
00:04:45 <thinrichs> #topic Newton 2
00:04:52 <thinrichs> Newton 2 is this week.
00:05:03 <thinrichs> I pushed the changes necessary for the release, updated launchpad, etc.
00:05:17 <thinrichs> So that should be mainly done, except for getting the patches merged
00:05:45 <thinrichs> Newton 3 is the end of August
00:05:50 <thinrichs> So about 6-7 weeks
00:06:07 <thinrichs> We released both the server and the client.
00:06:09 <thinrichs> Questions?
00:06:32 <thinrichs> #topic Barcelona
00:06:57 <thinrichs> A reminder that today in about 4 hours is your last chance to submit a talk for the summit
00:07:14 <thinrichs> Questions, comments?
00:07:34 <ekcs> are there client changes we’re looking to make in N?
00:07:48 <thinrichs> ekcs: nothing specific that I know of
00:08:07 <ekcs> ok.
00:08:18 <thinrichs> Except perhaps for masahito's change where datasource config will be updatable
00:08:35 <thinrichs> Are you still planning on that, masahito?
00:09:49 <masahito> on client side?
00:09:57 <thinrichs> Yep
00:10:34 <thinrichs> If you're adding API calls to the server, I assume we're adding client-side commands to match
00:10:53 <masahito> I'm not sure how we need to change client codes now, I'll check it by next meeting.
00:11:18 <thinrichs> I'd imagine we'd want people to be about to write something like...
00:11:26 <masahito> s/how/how much/
00:11:30 <thinrichs> openstack congress datasource config update …
00:12:11 <thinrichs> ekcs: you were asking because clients get frozen before the servers, right?
00:12:20 <thinrichs> so we'd need to get those changes in earlier
00:12:23 <masahito> ok
00:12:26 <ekcs> right.
00:13:04 <thinrichs> masahito: that's something to be aware of—both the API and the client changes need to merged earlier than the server changes
00:13:57 <thinrichs> Moving on…
00:14:02 <thinrichs> #topic Reno
00:14:38 <thinrichs> There's a new tool out (Reno) that's supposed to help us write release notes at we make the code changes
00:15:08 <thinrichs> So the idea is that when you make a change that should have a release note, you include in your change a file that describes the release note
00:15:26 <thinrichs> Then the release notes are automatically generated for the release
00:15:35 <thinrichs> At least, that's what I understand.
00:15:41 <thinrichs> Here's a thread on the ML about it.
00:15:41 <thinrichs> http://lists.openstack.org/pipermail/openstack-dev/2016-June/098335.html
00:16:22 <thinrichs> The way it works is that we add a tox target that we run manually whenever we want to add one of these files to a commit
00:16:50 <thinrichs> and the tox target generates a file that we edit and include in our change.
00:16:56 <thinrichs> Then we send our change to gerrit as usual
00:17:40 <thinrichs> It's something to look into
00:18:05 <thinrichs> That's all I have on that topic.  Anyone else?
00:19:29 <thinrichs> #topic Gating
00:19:40 <thinrichs> Haven't checked recently.  What's going on with the gate
00:19:44 <thinrichs> ?
00:20:03 <ekcs> really quick thing. gate-congress-new-arch-devstack timeouts a lot.
00:20:10 <ekcs> anyone know how to increase the time limit?
00:20:15 <ekcs> #link https://bugs.launchpad.net/congress/+bug/1599991
00:20:15 <openstack> Launchpad bug 1599991 in congress "tempest test timeouts" [Medium,New]
00:21:15 <thinrichs> ramineni: not off the top of my head.  do you know?
00:21:22 <ekcs> can comment on the bug later if we want to move on to saev time.
00:21:29 <ramineni_> no .. i have to check
00:21:50 <masahito> timeout on executing Tempest, right?
00:21:55 <ekcs> right.
00:22:44 <masahito> When I check dse2 I hit this.
00:22:49 <masahito> #link https://bugs.launchpad.net/congress/+bug/1600171
00:22:49 <openstack> Launchpad bug 1600171 in congress "Some API takes a long time before returning 404" [Low,New]
00:23:12 <masahito> It could cause the timeout issue.
00:23:53 <ekcs> interesting. good to know.
00:24:28 <ekcs> I think we can move on.
00:24:35 <thinrichs> ramineni: do you know if we're making calls to non-existent nodes in the tempest tests?
00:25:10 <ramineni_> thinrichs: no, i dont think so ..
00:25:11 <ekcs> we’re definitely checking for datasources that don’t yet exist in the HA test.
00:25:29 <ekcs> but that’s not the same thing.
00:25:48 <ramineni_> ekcs: right, that could also take time.. ill have a look
00:25:48 <thinrichs> Good to keep that in mind
00:26:44 <thinrichs> thanks ramineni
00:26:48 <thinrichs> moving on...
00:26:52 <thinrichs> #topic Status
00:27:03 <thinrichs> ekcs: want to give us a status update?
00:27:21 <ekcs> I added the HA tasks to launchpad.
00:27:22 <ekcs> #link https://blueprints.launchpad.net/congress/+spec/high-availability-design
00:27:23 <ekcs> #link https://bugs.launchpad.net/congress/+bugs?field.tag=haht
00:27:29 <ekcs> submitted patch allowing ExecutionDriver to handle exec requests from replicated PEs. #link https://review.openstack.org/#/c/339965/
00:27:36 <ekcs> Also digging and asking around to see if there is a way to (at least partially) automate the identification of blocking calls in Python code.
00:27:36 <ekcs> Similar to what I did manually here: #link https://review.openstack.org/#/c/335721/
00:27:37 <ekcs> Ideally, all new code could be annotated automatically to help us identify thread-safety issues in review. It can easily happen that I change a method A (say add an rpc call), which introduces blocking calls into a bunch of things that use method A through many layers of calls.
00:27:39 <ekcs> That’s all from me.
00:29:19 <thinrichs> Looks like you and masahito have been digging into that ExecutionDriver patch.  Anything to discuss?
00:30:48 <ekcs> I think we’re mostly on the same page. Anything you want to discuss, masahito ?
00:31:23 <masahito> Nothing to discuss for the patch.
00:32:19 <masahito> I'll quick review latest patch after this meeting.
00:32:46 <thinrichs> ekcs: it'd be cool to auto-identify new blocking issues
00:33:03 <thinrichs> let us know what you find.
00:33:22 <ekcs> thinrichs: ok. it’s theoretically not hard, but I don’t know if there is a good existing tool.
00:33:45 <thinrichs> As reviewers we should all be especially careful with API changes, since that seems to be the most obvious place for those
00:34:20 <thinrichs> masahito: want to do a status update?
00:34:28 <masahito> ok
00:34:50 <masahito> starting to pick up HAHT stuff.
00:35:01 <masahito> #link https://bugs.launchpad.net/congress/+bug/1600011
00:35:01 <openstack> Launchpad bug 1600011 in congress "HAHT - API routing to support replicated PE" [High,New] - Assigned to Masahito Muroi (muroi-masahito)
00:35:13 <masahito> and checked datasource manager dependency.
00:35:29 <masahito> #link https://review.openstack.org/#/c/339385/
00:35:39 <masahito> #link https://review.openstack.org/#/c/339386/1
00:36:10 <masahito> I noticed we already get rid of datasource manager in dse2
00:36:44 <masahito> b/c the tempest test in second patch succeed.
00:36:50 <masahito> that't from my side.
00:36:54 <thinrichs> Yep—datasource mgr functionality is part of dsenode now
00:37:21 <ekcs> masahito: would you be interested in picking up this piece? since you’re most experienced with Pacemaker.
00:37:27 <ekcs> #link https://bugs.launchpad.net/congress/+bug/1602853
00:37:27 <openstack> Launchpad bug 1602853 in congress "HAHT - Resource agents for Pacemaker integration" [High,New]
00:37:44 <ekcs> possibly instead of the API routing.
00:38:50 <masahito> no problem
00:39:20 <ekcs> awesomeness
00:39:39 <thinrichs> ekcs: are there other pieces of HAT you're looking to offload?
00:39:41 <masahito> Which do we prefer to documenting or implement a custom resource agent?
00:40:23 <ekcs> I think either would work fine? But i’d defer to you on that.
00:41:50 <masahito> I think documenting is simpler for user who is not familir with Congress, so I start from documenting.
00:41:50 <thinrichs> aimeeu: want to do a status update?
00:41:59 <aimeeu> sure
00:42:17 <aimeeu> I was out a few days last week and pulled onto other things Monday/Tuesday, but today I was finally able to stack successfully -- played around with congress CLI and REST API --  and started looking at this API bug
00:42:18 <aimeeu> #link https://bugs.launchpad.net/congress/+bug/1597243.
00:42:18 <openstack> Launchpad bug 1597243 in congress "Policy Detail API returns 500 status codes in dse2" [Undecided,In progress] - Assigned to Aimee Ukasick (aimeeu)
00:42:28 <aimeeu> I also ran into trouble with the Policy dashboard in Horizon and after talking to thinrichs, opened this bug:  #link https://bugs.launchpad.net/congress/+bug/1602837
00:42:28 <openstack> Launchpad bug 1602837 in congress "Policy UI (Horizon): Unable to get policies list (devstack)" [Undecided,New]
00:42:43 <aimeeu> ramineni: since you've done some work on the horizon plugin, do you have any idea why the error is happening?  I'm happy to work on the bug and would appreciate any suggestions.
00:43:16 <ramineni_> aimeeu: you are using devstack?
00:43:24 <aimeeu> ramineni: yes
00:44:16 <aimeeu> ramineni: if you have time to take a quick look and add to the bug report if you have any ideas, I'd appreciate it
00:44:19 <ramineni_> aimeeu: i think the problem is authentication with keystone v3 , need some changes in horizon code
00:44:53 <ramineni_> i thought f submitting patch before
00:45:35 <aimeeu> ramineni: I'll add your comment to the bug report, and assign the bug to myself. Thanks!
00:45:48 <aimeeu> that's all from me
00:46:04 <ramineni_> sure, you can check the authentication part in horizon code, that should support v3
00:47:05 <ramineni_> let me know if you need any help
00:47:17 <aimeeu> will do - thanks!
00:47:52 <thinrichs> aimeeu: we should definitely get our Horizon interface working before Newton3, so it'd be fantastic if you could make that happen!
00:48:17 <thinrichs> ramineni: want to give a status update?
00:48:35 <aimeeu> thinrichs: OK, I'll start work on that bug tomorrow.
00:49:10 <ramineni_> i have submitted the initial design for synchronizr, i tried to minimize the depedency between both as discussed
00:49:30 <ramineni_> i will incorporate comments today
00:49:56 <ramineni_> thats it from my side
00:50:42 <thinrichs> It seems that our consistency model for HA will rely heavily on the synchronizer, so it'll be good to get that in place
00:51:19 <thinrichs> Here's my status update…
00:52:01 <thinrichs> I've been experimenting with setting distributed_arch to True by default and ripping out all the old dse code.
00:52:25 <thinrichs> The code change is pretty much done.
00:52:49 <thinrichs> Now I'm looking at deleting the old tests and replacing them with tests from tests2
00:53:01 <thinrichs> No major hurdles yet.
00:53:13 <ekcs> great!
00:53:34 <masahito> great!
00:54:04 <thinrichs> We did a really nice job of making this pretty easy, at least in terms of the code changes.
00:54:14 <thinrichs> Tests aren't quite so nice, but those are hard to deal with anyway
00:54:34 <thinrichs> That's it from me
00:54:52 <thinrichs> 5 minutes left.
00:54:55 <thinrichs> #topic open discussion
00:54:57 <ekcs> thinrichs: on HA tasks (your earlier question), anything unassigned, people are welcome to pick up. My plan is to leave most things unassigned and assign them to myself as I finish previous tasks. but the places I most need others’ expertise would be the resource agent part we just talked about and the tempests parts.
00:54:59 <ekcs> #link https://bugs.launchpad.net/congress/+bug/1600021
00:54:59 <openstack> Launchpad bug 1600021 in congress "HAHT - tempest tests, basic" [Medium,New]
00:55:00 <ekcs> #link https://bugs.launchpad.net/congress/+bug/1600019
00:55:01 <openstack> Launchpad bug 1600019 in congress "HAHT - tempest tests, failures" [Medium,New]
00:55:01 <ekcs> And to a lesser extent push data persistence (sqlalchemy experience)
00:55:02 <ekcs> #link https://bugs.launchpad.net/congress/+bug/1600015
00:55:02 <openstack> Launchpad bug 1600015 in congress "HAHT - push data persistence" [Medium,New]
00:56:45 <ramineni_> ekcs: ill check how much we can test in tempest part
00:57:12 <ramineni_> ekcs: we need first synchronizer in place for this, i will check it up once that is done
00:57:22 <ekcs> yup. thanks!
00:58:01 <thinrichs> ekcs: That list of bugs is great!  They're the High priority ones, right?
00:58:03 <masahito> once I push RA stuff, I'll take push data one.
00:58:49 <thinrichs> I'll try to pick something up after I rip out dse
00:59:12 <ekcs> thinrichs: which list of bugs? the list includes bugs at diff priotities.
00:59:14 <ekcs> thanks guys.
01:00:03 <thinrichs> Times up for this week.  Thanks all!
01:04:34 <masahito> #endmeeting