00:01:52 <ekcs> #startmeeting CongressTeamMeeting
00:01:52 <ramineni_> Hi
00:01:52 <openstack> Meeting started Thu Oct 20 00:01:52 2016 UTC and is due to finish in 60 minutes.  The chair is ekcs. Information about MeetBot at http://wiki.debian.org/MeetBot.
00:01:53 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
00:01:56 <openstack> The meeting name has been set to 'congressteammeeting'
00:01:57 <thinrichs> Hi all
00:02:19 <masahito> hi
00:02:35 <ekcs> hi all! hope everyone’s doing well =)
00:02:49 <ekcs> today’s topcis are pretty basic.
00:02:51 <ekcs> - Barcelona sessions
00:02:52 <ekcs> - PTG
00:02:52 <ekcs> - open discussion / status
00:02:59 <ekcs> anything else we want to talk about?
00:03:02 <aimeeu> hi all
00:03:37 <thinrichs> ekcs: Sounds good to me
00:04:12 <ekcs> Cool let's dive in then.
00:04:16 <ekcs> #topic Barcelona sessions
00:04:21 <ekcs> Here's our design summit schedule: #link https://www.openstack.org/summit/barcelona-2016/summit-schedule/#day=2016-10-27&summit_types=10&tags=1235
00:04:56 <ekcs> I reached out to Vitrage PTL and they may join one of our first two sessions to talk about opportunities for congress and Vitrage services to leverage each other.
00:04:57 <ekcs> They also invited those of us interested to join their work session where congress *may* be a topic of discussion.
00:04:58 <ekcs> #link https://www.openstack.org/summit/barcelona-2016/summit-schedule/events/17118/vitrage-work-session
00:05:23 <ekcs> Any thoughts or comments?
00:07:58 <ekcs> So right now we are expecting Adam (quotas, multi-region), and Nate (Neutron fwaas) for the first session.
00:07:59 <ekcs> Bryan and Masahito for the second sessions as well as additional use case discussion.
00:08:00 <ekcs> Vitrage may participate in one of them.
00:08:32 <ekcs> Please let me know over email or whatever if you have someone else who may want to discuss something at these sessions.
00:08:49 <thinrichs> ekcs: did you ever confirm with Bryan?  I never heard back about my proposed 9:50 timeslot
00:09:35 <ekcs> thinrichs: oh I didn’t know it wasn’t confirmed. So that’s a maybe.
00:09:49 <thinrichs> He agreed to talk, but never at a specific time
00:10:05 <thinrichs> That's why he's under the "without confirmed time-slots" bit of the etherpad
00:10:10 <thinrichs> I'll ping him
00:10:33 <ekcs> thinrichs: ah okay. yea that’d be helpful. thanks!
00:11:06 <ekcs> I see that people have been adding to the work sesions etherpad: #link https://etherpad.openstack.org/p/congress-ocata-design-summit
00:12:00 <ekcs> which is great. hopefully we can have thought about things and ready to make progress when we meet at the summit =)
00:13:06 <ekcs> #action thinrichs confirm time with bryant_att
00:13:19 <ekcs> #action ekcs confirm time/presence with vitrage team
00:13:37 <ekcs> Here's an etherpad where we can note down sessions/talks that are potentially interesting to Congress contributors #link https://etherpad.openstack.org/p/congress-interests-barca
00:13:56 <ekcs> any more thoughts on barcelona?
00:14:25 <thinrichs> Anyone see the rock-climbing thread?  Anyone plan on going?
00:15:00 <aimeeu> I like my bones to remain unbroken, so no
00:15:04 <thinrichs> :)
00:15:29 <thinrichs> I was thinking about it, so if anyone else is interested we could go together.  Just offering
00:15:50 <thinrichs> Nothing else on barcelona for me then
00:15:56 <ekcs> oh cool. I haven’t seen it but it sounds interesting!
00:16:52 <masahito> thinrichs: rock-climbing in barcelona?
00:17:10 <thinrichs> #link https://www.mail-archive.com/openstack-dev@lists.openstack.org/msg94468.html
00:17:43 <thinrichs> masahito: yep.  indoor.
00:17:45 <thinrichs> Better link:
00:17:47 <thinrichs> #link http://lists.openstack.org/pipermail/openstack-dev/2016-October/105844.html
00:18:25 <thinrichs> Not sure if we need to decide now or if we can decide that night
00:19:38 <masahito> looks interesting. thanks.
00:20:14 <ekcs> yea. Looks like nobody has decided on time. I guess we will figure it out.
00:20:56 <ekcs> ok moving on then.
00:21:04 <ekcs> #topic PTG
00:21:43 <ekcs> based on feedback that everyone mostly thought work sessions at the PTG is a good idea if we can make it, and that most of us expect to be able to make it, we've indicated to the organizers that we expect to have work sessions at the PTG in Atlanta (Feb 20-24, 2017)
00:22:03 <ekcs> Please start working with your orgs and budgets to secure the trip if possible =)
00:22:06 <ekcs> Any thoughts?
00:23:12 <ekcs> I am looking forward to having longer sessions to do deeper design and work and discussion. As well as more opportunity to work and interface with other teams.
00:23:51 <njohnston> +1
00:23:55 <thinrichs> +1
00:25:41 <ekcs> In the FAQ, we expect the PTG to be scheduled in a way to allow a vertical team contributor to participate in a horizontal project and vice versa. I may think about having some involvement woth oslo or QA teams.
00:26:50 <ekcs> Ok great let’s move on.
00:26:55 <ekcs> #topic open discussion / status
00:27:21 <ekcs> anyone have other topics or status updates?
00:28:24 <masahito> no meeting in next week, right?
00:29:03 <ekcs> oh right almost forgot. let's cancel IRC meeting next week cuz we'll be at the summit.
00:29:27 <ekcs> not to mention it would be like at midnight or something haha.
00:30:11 <thinrichs> Had a couple of questions about changesets in gerritt
00:30:28 <thinrichs> The local HA tests…are they ready to go in?
00:30:29 <thinrichs> https://review.openstack.org/#/c/358927/
00:31:15 <thinrichs> There's a comment about changing bindep later.  Does that comment pertain to newton or master?
00:31:29 <ekcs> right I think it’s ready as soon as pep8 is cleaned up.
00:31:39 <thinrichs> I can fixup the pep8 if that helps
00:32:07 <ekcs> I tihnk it’s ready for master. and we can reduce the bindep.txt in future patch.
00:32:17 <thinrichs> Are we porting that to newton as well?
00:34:21 <ekcs> hmmm. i guess we could. may help to have more tests if we need to debug things and make changes.
00:34:24 <ekcs> any downside?
00:35:38 <thinrichs> Any bugfix that goes into newton really should pass those tests.
00:36:18 <ekcs> yup that makes sense.
00:36:36 <ekcs> I was more thinking openstack rule type of downsides. but seems like the right thing to do.
00:36:38 <thinrichs> I think the cost of trying to remember to backport that patch later in case we add a bugfix is higher than the cost of adding the change to stable
00:37:05 <thinrichs> Tests were one of the exceptions to the critical/security fix rule of thumb
00:37:23 <ekcs> yup that makes sense.
00:37:49 <thinrichs> Second question…
00:38:23 <thinrichs> This patch is targeted for stable/newton, but do we think it needs to go in?
00:38:25 <thinrichs> #link https://review.openstack.org/#/c/376166/
00:39:08 <thinrichs> Most of the change is renaming an internal variable.  Rest seems to be forcing names to not start with underscores.
00:40:11 <thinrichs> I'm not even sure the code-change matches the commit message (anymore)
00:40:43 <ekcs> My thinking was it may or may not go in to newton depending on whether we see issues.
00:41:10 <ekcs> in distributed deployment, if one deletes a DS and immediately adds again, it would fail. this patch fixes that.
00:41:28 <thinrichs> I must have missed that bit
00:41:33 <ramineni_> I think fix for the bug is one line change ..may be we can have separate patch to fix only that
00:41:37 <ekcs> but when we discussed it last meeting or earlier, we thought it’s not critical enough.
00:43:45 <ekcs> my inclination is to leave things alone and not merge it. not critical. but if we do end up needing to make the fix we can resurrect and do the minimal fix. what do you guys think?
00:43:55 <thinrichs> ramineni_: I like that idea—isolate the bug fix
00:44:13 <ekcs> +1 on isolate fix.
00:44:24 <ekcs> if we fix it.
00:44:45 <thinrichs> Anyone point me to the actual bug fix?  Still haven't found it.
00:44:46 <masahito> +1 for isolation idea.
00:45:12 <ekcs> I did not register a bug for it.
00:46:24 <thinrichs> I meant the line in that change that fixes something.  Just looked again and didn't see it.
00:47:16 <thinrichs> We can always finish this discussion in gerrit
00:47:31 <ramineni_> https://review.openstack.org/#/c/376166/1/congress/dse2/dse_node.py
00:48:14 <ramineni_> I think removing is_valid_service check ..which does in mem
00:48:31 <ramineni_> ekcs: right?
00:48:53 <ekcs> at right exactly. I was having trouble remembering and finding it too. thanks
00:49:01 <ekcs> line 721
00:50:30 <thinrichs> So if there's a synchronization race, and someone deletes datasource named D and then someone adds datasource named D, and the synchronizer hasn't run in between, then we would reject a valid datasource request
00:51:06 <thinrichs> I could imagine that happening easily in HA if the user isn't pinned to an API node by the LB
00:51:16 <thinrichs> Then again, we recommend pinning anyway.
00:51:32 <ekcs> correct. but pinning doesn’t matter in this case.
00:51:52 <thinrichs> Are we not deleting the DS out of memory on the node that accepts the delete-DS call?
00:52:01 <thinrichs> If so, I think we should merge that into newton
00:52:27 <thinrichs> Got to run for tonight.  Thanks all!
00:52:45 <ekcs> later thinrichs. let’s wrap up and we can finish that discussion on gerrit.
00:52:53 <ekcs> What would be a good way for us to optionally communicate and coordinate with each other at the summit? Email? Messaging app?
00:53:41 <ramineni_> Messaging app??
00:54:01 <aimeeu> IRC app?
00:54:39 <ekcs> aimeeu: do you know a good one that takes care of offline messages?
00:54:58 <ekcs> singe at the summit we’ll probably be disconnected a lot.
00:55:02 <aimeeu> irccloud has web and mobile apps
00:55:08 <ramineni_> If works WhatsApp should be good .. ??
00:55:28 <ramineni_> Not sure everyone uses though
00:55:29 <aimeeu> the free version has a trial with 7 days history
00:56:16 <aimeeu> #link http://irccloud.com
00:56:58 <aimeeu> the Women of  OpenStack use GroupMe
00:58:13 <ekcs> haha so many things.
00:58:32 <ekcs> maybe we should give irc a try singe we already use it?
00:58:49 <aimeeu> +1
00:59:02 <ramineni_> Sounds good
00:59:36 <ekcs> well time’s up. we can finish the discussion over ML for those interested. See you all soon!
00:59:47 <aimeeu> See you next week!
00:59:52 <masahito> see you next weekl
01:00:03 <ekcs> #endmeeting