00:00:49 <ekcs> #startmeeting congressteammeeting
00:00:50 <openstack> Meeting started Thu May 18 00:00:49 2017 UTC and is due to finish in 60 minutes.  The chair is ekcs. Information about MeetBot at http://wiki.debian.org/MeetBot.
00:00:51 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
00:00:53 <openstack> The meeting name has been set to 'congressteammeeting'
00:01:02 <ramineni_> ekcs: hi
00:01:09 <ekcs> Hi all! welcome back to another meeting.
00:01:11 <ekcs> hi ramineni_ !
00:01:32 <masahito> hi
00:01:42 <ekcs> as usual, collaborative set of topics are here; https://etherpad.openstack.org/p/congress-meeting-topics
00:03:38 <ekcs> hi masahito !
00:05:02 <ekcs> ok let's dive in then.
00:05:48 <ekcs> #topic gate failures
00:06:19 <ekcs> ramineni_ has resolved/worked around the failures from last week. Big thanks!
00:06:47 <ekcs> seems like we may have a new failure though.
00:07:07 <ramineni_> Nova client removed the floating ips list method from their code ..so we are seeing error
00:07:56 <ramineni_> It's deprecated since ocata I guess .. They removed completely now
00:08:46 <ekcs> Got it! I had started looking at it today but didn't get very far.
00:08:56 <ekcs> ramineni_, do you know if there is an easy fix?
00:09:15 <ramineni_> I raised a patch
00:09:43 <ekcs> ah I see this one: #link https://review.openstack.org/#/c/465020/
00:09:43 <patchbot> patch 465020 - congress - Remove floating_IPs from nova datasource driver
00:09:57 <ramineni_> I removed it from our code for now .. Do you see a better way
00:10:05 <ramineni_> ?
00:10:23 <ramineni_> ekcs: yes
00:11:07 <ekcs> Hmmm. Do you have any reference to the deprecation and how they suggest people get that information?
00:11:38 <ramineni_> I added link to release notes
00:11:47 <ramineni_> In commit message
00:12:08 <masahito> warning messages?
00:12:09 <ekcs> great. also this reference: #link https://specs.openstack.org/openstack/nova-specs/specs/pike/approved/deprecate-multinic-proxy-api.html
00:14:10 <ramineni_> masahito: do you mean to add warning message it's deprecated in update datasource s method
00:14:40 <masahito> ramineni_: yes.
00:15:19 <masahito> usually, user can't read log messages, but admin could check and notify it to user.
00:15:51 <ramineni_> Ya , I can add that ..
00:16:08 <masahito> I don't have nice ideas to notify the table is deprecated of user...
00:16:14 <ramineni_> But it's completely removed now rather than deprecated
00:16:18 <ekcs> Not sure what the best approach is. 1) just remove it. 2) remove the data but keep a dummy column so that policies with positional args that do not use the column anyway are not broken.
00:17:17 <masahito> IIRC, congress can keep other tables that refer to the floating_ip table.
00:17:38 <ekcs> 3) keep the call in place, so it works with older nova/client and on newer nova log error message and continue.
00:18:40 <ekcs> masahito, I don't totally understand what you mean.
00:19:05 <masahito> for instance,
00:19:39 <masahito> admin has a rule like x(y):- nova:floating_ip(z)
00:19:54 <masahito> then upgrade it without updating the rule.
00:20:25 <masahito> oops, x(y) :- nova:floating_ip(y)
00:20:38 <ramineni_> Upgrade it to neutron floating ips?
00:21:09 <masahito> the table 'x' has no row because of no row in nova:floating_ip(z)
00:21:19 <masahito> ramineni_: yes
00:22:42 <masahito> of course, the rule is needed to upgrade before upgrading congress.
00:23:02 <ekcs> I see. so masahito you're saying we should keep the table just leave it empty.
00:23:49 <masahito> if congress doesn't take care the situation, it's ok to completely remove the table.
00:24:39 <ekcs> oh. so you're saying we should change the nova_driver to get the floating IPs from neutron.
00:24:48 <masahito> ekcs: leave it empty or write a message to notify the user.
00:25:14 <ramineni_> masahito: ok , got it
00:25:35 <ekcs> haha ok I'm still not sure what the proposal is.
00:26:09 <ekcs> 1. remove table. 2. leave empty table. 3. populate table with data from neutron.
00:26:09 <masahito> ekcs: i don't think it's good to get info from neutron_driver....
00:26:12 <ekcs> is it one of those?
00:26:28 <ekcs> I don't either. Just trying to understand which one you're saying we should do.
00:27:32 <ramineni_> We should add a message this table is deprecated .. neutron table should be used further for floating ip data?
00:28:13 <ramineni_> And leave the table empty
00:29:28 <ekcs> I prefer removing the table completely. It's better to break things and force admin to fix immediately, rather than potentially execute very unintended effects.
00:30:07 <ekcs> say execute[action(x)] :- server(x, y), not nova:floating_ips(y)
00:30:19 <ekcs> now all of a sudden many things are executed.
00:30:31 <ekcs> when someone upgrades congress.
00:31:15 <ekcs> what do you guys think?
00:31:43 <ramineni_> ekcs: sounds good to me .. it will be clean
00:32:17 <masahito> I don't have any objection to remove the table.
00:32:59 <ekcs> cool so basically what ramineni_ already did in the patch. great!
00:33:23 <ekcs> ok let's move on then. that took a while haha.
00:33:28 <ekcs> #topic policy library
00:33:47 <ekcs> good discussion last time about policy library. and whether to use templates, etc.
00:33:56 <ekcs> any further thoughts on it to share here?
00:34:47 <ramineni_> ekcs: I'm good with Tim's suggestion to proceed without template s for now.
00:35:14 <ramineni_> Later we can see if we can explore that templating mechanism
00:35:52 <ekcs> ramineni_, got it!
00:36:17 <ekcs> masahito, do you have anything to add about the policy library spec?
00:36:43 <masahito> sorry, I've not check well yet....
00:37:15 <masahito> I try to check it today, so if no comment from my side. please go ahead
00:37:40 <ekcs> no problem. ok thanks!
00:38:11 <ekcs> ok next topic then.
00:38:14 <ekcs> #topic PTG
00:38:42 <ekcs> A few weeks ago we mentioned we need to start deciding whether to do meetings at the next PTG.
00:39:20 <ekcs> expected: September 11-15; Location: Denver, CO
00:39:26 <ekcs> We need to have a decision by this Sunday at the latest.
00:40:17 <ekcs> So basically I just need to know from folks: 1) whether you think we should do sessions at the PTG, and 2) if we do, whether you expect to be able to make it.
00:40:52 <ekcs> Please feel free to discuss it here and/or email me later (I sent an email to ML)
00:42:34 <masahito> if we have, I'm ok to go there
00:43:05 <ramineni_> ekcs: ok, I think I should be able to attend
00:43:35 <ramineni_> If we are planning to have
00:43:48 <masahito> but if most of us can't go, I don't think it's need to have meeting.
00:44:27 <masahito> s/meeting/ meetings at PTG/
00:44:57 <ekcs> ok so masahito I'm understanding that as: if most people can go, it'd be good to do it, otherwise, better to not do it. is that accurate?
00:45:16 <masahito> ekcs: yes
00:46:13 <ekcs> ok perfect. ramineni_ any more thoughts on whether you think we should have it or not? or email me later?
00:46:49 <ramineni_> Agree with masahito
00:47:01 <ekcs> ok great. thanks!
00:47:12 <ekcs> let's move on then.
00:47:21 <ekcs> #topic patches
00:47:31 <ekcs> any other patches that'd be helpful to talk about here?
00:47:45 <ramineni_> When is pike 2
00:48:24 <ramineni_> ekcs: have a question on congress-dashboard release
00:48:35 <ekcs> https://releases.openstack.org/pike/schedule.html#p-2
00:49:03 <ekcs> week of june 5.
00:49:10 <ekcs> wow that's right around the corner.
00:49:40 <ekcs> ramineni_, what's the question?
00:50:11 <ramineni_> Ok .. how new projects are released .. should we do it for pike2
00:50:24 <ramineni_> For congress- dashboard
00:50:49 <ekcs> hmm let me find out from the release team.
00:51:03 <ramineni_> Ok
00:51:14 <ekcs> #action ekcs find out whether to release congress-dashboard pike-2
00:52:19 <ekcs> on patches, here's a piece needed for policy library: #link https://review.openstack.org/#/c/460348/
00:52:20 <patchbot> patch 460348 - congress - support transactional insert of multiple rules
00:52:28 <ekcs> pls review when you have a chance ;)
00:52:42 <ekcs> anything else to bring up on patches?
00:54:22 <ekcs> ok let's move on then!
00:54:27 <ekcs> #topic open discussion
00:54:37 <ekcs> any last things?
00:55:20 <ramineni_> None from my side
00:55:28 <ekcs> masahito, how was the summit for you? anything interesting?
00:56:12 <masahito> I was busy with another project's stuffs....
00:56:17 <masahito> at the summit
00:56:43 <ekcs> of course. I didn't mean necessarily congress specific =)
00:56:58 <masahito> =)
00:57:02 <ekcs> I wasn't able to make it to the rabbitMQ talk, but hope it went well.
00:57:16 <ekcs> I mostly went to different forum sessions.
00:57:54 <ekcs> It seems tempest plugin will likely be a Queens goal. https://review.openstack.org/#/c/369749/
00:57:55 <patchbot> patch 369749 - governance - Add Queens goal split out tempest plugins
00:58:14 <ekcs> Based on how the discussion went at the forum.
00:59:27 <ekcs> we also discussed deprecating postgreSQL, which would make many things easier for devs. But some ops really depend on postgres, so I think official deprecation is off the table at the moment.
01:00:25 <ekcs> ok well time's up.
01:00:42 <ekcs> bye and have a great week/weekend!
01:00:46 <ekcs> #endmeeting