00:01:14 <thinrichs> #startmeeting CongressTeamMeeting
00:01:15 <openstack> Meeting started Thu Mar 24 00:01:14 2016 UTC and is due to finish in 60 minutes.  The chair is thinrichs. Information about MeetBot at http://wiki.debian.org/MeetBot.
00:01:17 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
00:01:19 <openstack> The meeting name has been set to 'congressteammeeting'
00:01:26 <ramineni_> thinrichs: hi
00:01:40 <tsandall> hello
00:01:48 <masahito> hi
00:02:57 <thinrichs> Agenda this week is short: discuss Mitaka release
00:02:59 <thinrichs> Anything else?
00:04:08 <thinrichs> #topic Mitaka
00:04:17 <thinrichs> Monday/Tuesday we cut RC1
00:04:40 <thinrichs> That means (i) master is open for non-mitaka changes and ...
00:04:55 <thinrichs> (ii) any change we want to go into Mitaka we need to target to stable/mitaka
00:05:21 <thinrichs> If you're submitting a change to mitaka, first submit it to master and then to stable/mitaka.
00:05:32 <thinrichs> I've done this for a couple of the changes that have merged recently.
00:05:56 <thinrichs> From the web UI of gerrit, you can Cherry Pick a change and enter stable/mitaka as the branch
00:06:16 <thinrichs> or you can submit to stable/mitaka from git in the usual way.
00:06:25 <thinrichs> Any questions?
00:06:34 <ekcs> should we do that before or after reviews?
00:06:49 <ekcs> cherrypick for mitaka that is.
00:06:51 <thinrichs> ekcs: I'd probably wait to submit to stable/mitaka until after you get reviews
00:07:12 <thinrichs> Once you get signed off for master, then cherry-pick to stable/mitaka.
00:07:19 <thinrichs> Just saves you work.
00:08:08 <thinrichs> Anything else?
00:08:12 <ekcs> got it.
00:09:41 <thinrichs> Next I think we do some status updates.  Are there bugs you know about that you're trying to fix by Mitaka?
00:09:44 <thinrichs> #topic status
00:09:51 <thinrichs> ramineni_: want to start?
00:10:12 <ramineni_> thinrichs: ok
00:11:23 <ramineni_> thinrichs: all the tempest tests in our repo are enabled now, murano is the last one
00:11:40 <ramineni_> thinrichs: and some small things i have found during testing, have fixed that
00:12:00 <ramineni_> thinrichs: i didnt find any critical bugs :)
00:12:09 <thinrichs> Do we want to add the murano test to mitaka?
00:12:51 <ramineni_> thinrichs: we can add, but we need project-config change to be in first
00:13:10 <ramineni_> thinrichs: https://review.openstack.org/#/c/296380
00:13:23 <ramineni_> thinrichs: after that i will add
00:13:59 <ramineni_> thinrichs: murano service is not enabled as of now in devstack
00:14:03 <thinrichs> Are you submitting that change to the stable/mitaka branch as well?
00:14:48 <ramineni_> thinrichs: ya, its a common job for all branches
00:15:02 <ramineni_> thinrichs: so, it works for mitaka as well
00:15:26 <thinrichs> ramineni_: understood.  That strikes me as surprising that there's 1 job for all branches.  Do you know why?
00:16:11 <ramineni_> thinrichs: ya :( , infra guys didnt agree for seperate job last time we submitted, remember?
00:16:25 <ramineni_> thinrichs: so, finally settled for one job
00:16:42 <thinrichs> ramineni_: right—not saying we should have separate jobs.  I think you're doing the right thing.
00:17:34 <thinrichs> ramineni_: It was more a question about why when they're running gate on stable/mitaka for congress that they don't pull out the project config for congress from the stable/mitaka branch of project-config
00:18:22 <ramineni_> thinrichs: project-config doesnt have branches
00:18:27 <ramineni_> thinrichs: only master
00:18:39 <thinrichs> ramineni_: I see—simple answer.
00:19:14 <ramineni_> thinrichs: :)
00:19:15 <thinrichs> ramineni_: anything else?
00:19:25 <ramineni_> thinrichs: no, im done
00:19:34 <thinrichs> masahito: want to go next?
00:19:41 <masahito> ok
00:20:47 <masahito> I took a look about the error about push driver isn't listed in datasource list.
00:21:08 <masahito> but I can't reproduce it in local.
00:22:08 <masahito> That's it from my side.
00:22:10 <thinrichs> It wasn't that it wasn't shown, it was that there was an internal error when listing it.
00:22:28 <thinrichs> Did anyone else try the push driver?
00:22:53 <thinrichs> I wonder if I had an older client.
00:23:57 <ekcs> thinrichs: haven’t tried.
00:24:11 <thinrichs> Here's the mail I sent...
00:24:13 <thinrichs> #link http://lists.openstack.org/pipermail/openstack-dev/2016-March/089737.html
00:24:29 <masahito> thinrichs: I'll check it once again.
00:25:03 <thinrichs> masahito: could someone else try it out as well?
00:25:30 <ekcs> thinrichs: i’ll give it a try.
00:25:47 <thinrichs> ekcs: thanks.  Maybe I'm crazy, but I hit 2 different problems.
00:25:55 <thinrichs> One seemed to be the server, the other the client.
00:26:52 <thinrichs> I think it's good to have a 3rd person in the loop.
00:28:20 <masahito> now I notice that the cause of the error of second one would be URL.
00:28:25 <thinrichs> ekcs: if you can take a look soon, I should be available for followup discussions this week.
00:28:37 <ekcs> thinrichs: ok.
00:28:58 <thinrichs> masahito: that might be the problem.
00:29:16 <masahito> datasource API model seems to accept /v1/<datasource-*id*>/table/<table-id>/rows ....
00:29:28 <thinrichs> masahito: I think I pieced together the email though from several different commands, so it could be a problem with my email as well.
00:30:07 <thinrichs> masahito: are you saying we can only use the UUID of the datasource for the push?
00:30:12 <thinrichs> I know I didn't try that.
00:30:51 <masahito> thinrichs: I think it's for all datasource. not only for push.
00:30:55 <thinrichs> I'll try that and see if I have better luck.
00:31:13 <masahito> I'll also try it.
00:31:22 <thinrichs> masahito: sorry—that totally makes sense.  The client translates from name to ID and the server only expects an ID.
00:31:36 <thinrichs> As an aside, I think the server ought to accept name or ID in all cases.
00:31:46 <thinrichs> since both are unambiguous.
00:31:59 <masahito> thinrichs: right.
00:32:25 <thinrichs> Assuming we get it working, which driver should we actually merge?
00:32:27 <masahito> thinrichs: do you think it's critical bug for RC?
00:33:09 <thinrichs> masahito: we've told people that they can push data into Congress in Mitaka.  So it'd be nice to have at least 1 driver merged.
00:33:29 <thinrichs> The downside to the DoctorDriver is that it has a fixed schema.
00:34:16 <thinrichs> I was trying to put together a version that let people push any data in any schema that want as long as the schema just includes tables
00:34:31 <thinrichs> (so no embedded structure that we'd need to translate into tables)
00:35:19 <thinrichs> The schema could then be computed at any point in time as just the collection of tables that have been pushed in.
00:35:52 <thinrichs> Thoughts?
00:36:21 <masahito> I think it's good as sample driver.
00:36:24 <ekcs> sounds good.
00:36:47 <masahito> I push DoctorDriver in next release.
00:37:20 <thinrichs> DoctorDriver in the next release makes sense
00:38:01 <masahito> And if technically possible I'll change DoctorDriver following the translation.
00:38:05 <thinrichs> Here's the sample driver I started:
00:38:10 <thinrichs> https://review.openstack.org/#/c/294348/
00:38:52 <thinrichs> masahito: once we verify that everything is working as we'd expect, do you have time to finish off the sample driver I started?
00:40:17 <masahito> it means merge it? or fix some comments?
00:41:20 <masahito> of cause, both are ok.
00:41:29 <thinrichs> masahito: I never got it working, so I'd imagine there's some work to do, and then we'll want to test it, and then get it reviewed, and merge
00:41:48 <masahito> thinrichs: got it.
00:42:26 <masahito> i'll do that
00:42:58 <thinrichs> masahito: I overrode 1 or 2 methods from the base class so we could avoid handing in translators.  Otherwise it should be straightforward.
00:43:02 <thinrichs> masahito: thanks!
00:43:26 <thinrichs> ekcs: how is your bug finding/squashing going?  I saw a couple fixes of late.
00:43:39 <ekcs> 1. Here's a critical fix for Mitaka. The way I did antlr3 py2vs3 selection broke the stand-alone install. https://review.openstack.org/#/c/295621/
00:43:41 <ekcs> 2. Still not sure if this is a bug or not. I haven't worked much more on it since the last comment. https://bugs.launchpad.net/congress/+bug/1559362
00:43:41 <openstack> Launchpad bug 1559362 in congress "heat ERROR: Authentication failed: Authentication required" [Undecided,New]
00:43:41 <ekcs> 3. I've been working on stand-alone install instructions. With this patch, it should be a working set of instructions. https://review.openstack.org/#/c/296826/
00:45:19 <thinrichs> For (2) I wonder if the token that the client is getting from Keystone is timing out.
00:45:58 <ekcs> So would that be a heat bug?
00:46:23 <thinrichs> Not sure.  Maybe there's just a way to tell the client to refresh its token.
00:46:38 <thinrichs> Maybe all the other services catch that error automatically and refresh the token.
00:46:49 <ekcs> hmm. ok.
00:47:06 <thinrichs> It's probably worth asking the Heat team if that can happen and if so how to deal with it.
00:47:15 <ekcs> ok
00:47:31 <ekcs> would mailing list be a good place?
00:47:42 <thinrichs> I'd think so.  Or the heat IRC.
00:47:48 <ekcs> great.
00:47:57 <thinrichs> I'd probably start with the ML and if you get no response in a day ping someone on IRC.
00:48:35 <ekcs> k.
00:48:58 <thinrichs> ekcs: nice work going through the standalone instructions.  I know bryan_att had trouble getting us installed.
00:49:26 <ekcs> yup.
00:49:46 <thinrichs> Let's open it up for discussion the last few minutes.
00:49:49 <thinrichs> #topic open discussion
00:51:07 <ramineni_> thinrichs: you were asking about stable/liberty release before, is it done?
00:52:07 <thinrichs> ramineni_: I'm not remembering right now.  Was I supposed to cut another release?
00:54:08 <ramineni_> thinrichs: ya, you mailed before .. dont remeber exactly when
00:54:42 <thinrichs> I'm looking at the changes that have gone in since liberty's 2.0.0 release
00:54:42 <ramineni_> thinrichs: but never mind , if we dont have any plans to release
00:55:09 <thinrichs> I remember that being on the list of things to do; I just don't remember why.
00:55:35 <thinrichs> If we have enough changes since 2.0.0 and we don't expect any more anytime soon, it makes sense to cut the release.
00:56:48 <thinrichs> Yeah—we should cut a release on liberty.  There are some bug fixes in there.
00:56:58 <ramineni_> thinrichs: ok, if we plan to cut a release, ill backport some neutron tests which are stabe now, last time i disabled them
00:57:20 <thinrichs> ramineni_: ok.  Let me know when you're done, and I'll cut the release.
00:57:34 <ramineni_> thinrichs: ok, sure
00:57:50 <ekcs> For Mitaka, does it make sense to prepare a set of python3 Congress install instructions? Or not really needed at this point?
00:58:19 <thinrichs> Are the instructions different for python2 and python3?
00:59:24 <ekcs> very similar. pip3 for pip, python3 for python. Other than that it’s the need to use pymysql because straight mysqldb python client doesn’t support 3
00:59:44 <ekcs> install and use pymysql
01:00:09 <thinrichs> I worry about adding a whole new set of install instructions this late in the release.
01:00:43 <ekcs> ok. I don’t really think people are going to miss it really.
01:00:55 <ekcs> 2.7 is still standard.
01:00:59 <ekcs> for deployment.
01:01:10 <thinrichs> Let's add those instructions to master though.
01:01:22 <ekcs> ok.
01:01:23 <thinrichs> Maybe even just give the python3 instructions
01:01:32 <thinrichs> Or at least think about it.
01:01:37 <thinrichs> Out of time.
01:01:40 <thinrichs> Thanks all!
01:01:43 <thinrichs> #endmeeting