00:01:46 <ekcs> #startmeeting congressteammeeting
00:01:47 <openstack> Meeting started Thu Feb  9 00:01:46 2017 UTC and is due to finish in 60 minutes.  The chair is ekcs. Information about MeetBot at http://wiki.debian.org/MeetBot.
00:01:48 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
00:01:50 <openstack> The meeting name has been set to 'congressteammeeting'
00:02:11 <ekcs> Hi all. hope everyone’s having a great week!
00:02:26 <ekcs> here are the topics I have for today. anything else on our minds?
00:02:30 <ekcs> topics:
00:02:32 <ekcs> - ocata bugs
00:02:32 <ekcs> - PTG topics discussion
00:02:33 <ekcs> - PTG social planning
00:02:41 <masahito> hi
00:02:47 <ekcs> - stable/newton gate issue
00:04:22 <ekcs> ok let’s get startedthen =)
00:04:31 <ekcs> #topic ocata bugs
00:04:47 <ekcs> are there any new bugs we’d like to talk about?
00:06:08 <thinrichs> I ran into one .. the flag delete-datasources doesn't work.  Let me dig up the bug number
00:06:39 <thinrichs> https://bugs.launchpad.net/congress/+bug/1663057
00:06:39 <openstack> Launchpad bug 1663057 in congress "Auto-deletion of datasources fails" [Undecided,New]
00:06:55 <thinrichs> There might be a related bug.
00:07:20 <thinrichs> If ceilometer isn't reachable, does the Ceilometer driver fail to initialize?  That seemed to be what happened to me.
00:07:44 <thinrichs> Example in that bug report
00:09:22 <ekcs> got it. so it seems like the deletion is requring the datasource be reachable?
00:09:51 <thinrichs> I don't think so.  Perhaps the writeup was misleading.
00:09:57 <thinrichs> I think there might be 2 distinct issues.
00:10:17 <thinrichs> 1. Supplying the delete-… flag on the command line references code that has been moved.
00:10:59 <ramineni_> thinrichs: I think bug seems to be because delete datasource is moved and not update in the function
00:11:16 <thinrichs> 2. The Congress ceilometer driver seems to need to talk to Ceilometer when initializing.  If it can't Congress doesn't start.
00:11:30 <thinrichs> ramineni_: agreed for #1
00:12:02 <ramineni_> Will add quick fix for that
00:12:13 <thinrichs> ramineni_: great!
00:12:32 <ekcs> got it.
00:12:35 <thinrichs> Does anyone know if the ceilometer driver connects to ceilometer on init?
00:12:47 <thinrichs> We probably don't want congress to fail to start b/c it can't reach one of the other services.
00:12:56 <ramineni_> It doesn't I suppose
00:13:08 <ekcs> so 1. the deletion flag implementation needs some fixing up.
00:13:20 <ramineni_> We just create client object
00:13:36 <ekcs> 2. ceilometer driver fails to initialize if ceilometer not responsive. probably doesn’t have to be that way.
00:14:14 <ekcs> agreed.
00:14:30 <masahito> agreed
00:14:44 <thinrichs> If either of those 2 issues were fixed, I could have worked thru the problem, but since both were there I had to manually edit the database.
00:15:05 <ramineni_> Isn't the error because aodh is not started yet
00:15:43 <thinrichs> ramineni_: You're right.  Not sure why I was thinking it was ceilometer.
00:16:41 <thinrichs> The only other bug I saw was Eric's 500 on a nova delete.  I've got a devstack that won't let me delete the 'nova' datasource.
00:16:49 <ramineni_> But it's happening with latest code ..I remember catching that exception and just return ing
00:16:54 <ramineni_> ??
00:17:27 <thinrichs> I just created a new devstack this afternoon, so it should be from the tip of master.
00:17:57 <ramineni_> Ok
00:18:27 <ekcs> yea. the nova delete bug is here.
00:18:29 <ekcs> https://bugs.launchpad.net/congress/+bug/1662809
00:18:29 <openstack> Launchpad bug 1662809 in congress "500 error on nova datasource delete" [Undecided,New]
00:18:33 <ekcs> #link https://bugs.launchpad.net/congress/+bug/1662809
00:18:59 <ekcs> i’ve narrowed down on how to reproduce it.
00:19:09 <ekcs> three things needed:
00:19:31 <ekcs> 1. a rule that mentions the datasource in head and body.
00:20:00 <ekcs> 2. the congress instance has to be “newly started”. if certain things take place, the error no longer occurs.
00:20:57 <ekcs> quite strange.
00:21:10 <thinrichs> What's the 3rd thing?
00:21:31 <ekcs> oops I guess just 2
00:21:48 <thinrichs> Know what causes "newly started" to be false?
00:22:17 <thinrichs> I went thru that script of yours once (kind of) manually, then deleted the policies and new datasources, and ran the script again and saw the bug.
00:22:19 <ekcs> thinrichs: I know i said over email maybe this is the cause of the depedency graph issue, but it could also be the dependence graph issue already exists and causes the 500 here.
00:24:03 <ekcs> thinrichs: actually I didn’t say it quite right.
00:26:14 <ekcs> here’s what happens.
00:26:15 <thinrichs> Hmmm…interesting wrinkle.
00:26:29 <thinrichs> Seems the nova datasource delete command returned 404 for me, but then deleted the datasource.
00:27:00 <ekcs> if I experience the error, then delete and create policy1 then run the same script again that generated the error the first time, error does not occur..
00:27:21 <thinrichs> Interesting
00:28:06 <ekcs> but if i create and delete right from the start of congress, error still happens. quite strange.
00:28:32 <ekcs> anyway. just want to mention it. not going to solve it here probably.
00:29:18 <ekcs> pretty important bug to fix because it is related to another issue that stops all rule insertions.
00:30:07 <ekcs> thinrichs: I think it returns 404 because the delete went through the DB but didn’t finish in congress. so when you run delete again after a DS sync, the ds wasn’t there any more.
00:30:22 <ekcs> If I add the DS again and delete again it’s back to the 500 error.
00:30:28 <thinrichs> ok
00:31:11 <ekcs> well let’s move on then.
00:31:21 <ekcs> unless someone wants to talk about another bug.
00:32:26 <ekcs> #topic PTG topics discussion
00:32:50 <ekcs> Thanks all for adding more thoughts to the etherpad. $link https://etherpad.openstack.org/p/congress-ptg-pike
00:32:56 <ekcs> #link https://etherpad.openstack.org/p/congress-ptg-pike
00:33:19 <ekcs> is there anything we want to quickly discuss/clarify here?
00:33:25 <ekcs> (may not be)
00:34:30 <ekcs> I’m wondering what is meant by this:
00:34:35 <ekcs> parallel enforcement: policy engine calls datasource driver serially when some actions needed.
00:34:52 <ekcs> which is the desired behavior? parallel or serial?
00:35:10 <masahito> it's an optional feature.
00:36:12 <masahito> I suppose congress has new flag, parallel enforcement or something like this.
00:36:16 <ekcs> masahito: got it. so which behavior is it that we want to consider adding? parallel or serial?
00:36:26 <masahito> parallel
00:36:46 <ekcs> for performance/responsiveness reasons?
00:36:58 <masahito> I'll add an example rule in the enforcement.
00:37:08 <masahito> ekcs: yes
00:37:20 <ekcs> masahito: got it!
00:37:49 <masahito> s/in the enforcement/in the etherpad/
00:38:31 <ekcs> so why don’t we continue adding our thoughts to the etherpad (if there is more), and I’ll send out on ML a proposed rough schedule that we can finalize next IRC?
00:38:42 <ekcs> based on all the topics in the pad.
00:39:13 <thinrichs> Sounds good
00:39:31 <ekcs> ok moving on then if there’s nothing else on this topic atm.
00:40:02 <ekcs> #topic PTG social
00:40:19 <ekcs> thinrichs sent out an email regarding this.
00:40:25 <ekcs> thinrichs: you wanna talk more about it?
00:40:39 <thinrichs> Here's the link
00:40:40 <thinrichs> http://lists.openstack.org/pipermail/openstack-dev/2017-February/111762.html
00:41:11 <thinrichs> I found 2 vegetarian-friendly restaurants for dinner (one Ethiopian, one American), both highly rated on trip advisor.
00:41:40 <thinrichs> I also found 2 outings: the aquarium and the botanical gardens
00:41:57 <thinrichs> We could do a combo of dinner and outing, or just 1 of them.
00:42:11 <ekcs> I’ve heard good things about the aquariam.
00:42:34 <thinrichs> Other preferences?
00:42:43 <thinrichs> Did we ever pick a day?  Wed night or Thu night?
00:43:15 <masahito> thinrichs: thank you! all of your choices looks great!
00:43:55 <thinrichs> Thanks!  I'm happy to do it.
00:43:58 <ekcs> I don’t think we picked a day.
00:45:05 <ramineni_> thinrichs: thanks for the options .. looks great
00:45:33 <ekcs> aquarium i heard is really great, but i’m not sure how much time people want to dedicate to it. if we do a quick mid-day trip, it feels like a waste of the $44 ticket.
00:46:00 <thinrichs> And we can always choose to do the aquarium / botanical gardens on the spur of the moment when we're there.
00:46:20 <thinrichs> Maybe we should choose a restaurant and a day and make reservations
00:46:38 <ekcs> great let’s do that.
00:46:54 <thinrichs> Any preferences for days or Ethiopian/American?
00:47:14 <thinrichs> If not, I can pick one when making reservations.
00:47:28 <ekcs> both restaurants seem good to me. no preference.
00:47:48 <ramineni_> Thursday should be good
00:47:49 <ekcs> I slightly prefer thursday but not by much.
00:48:36 <ramineni_> Ethiopian looks far right
00:48:36 <thinrichs> masahito, aimeeu: preference on day?
00:48:54 <thinrichs> Ethiopian is a little farther away (20min).
00:48:54 <masahito> I prefer Thursday
00:48:54 <aimeeu> no preference on day or restaurant
00:49:05 <thinrichs> Thursday it is then.
00:49:33 <thinrichs> I'd lean toward the american place, since it's closer.
00:50:16 <ekcs> sounds good to me.
00:50:35 <ramineni_> Sounds good ..
00:50:37 <masahito> me too
00:51:03 <thinrichs> Okay.  I'll make the reservation for Thu night at the American place.  Say 7p?
00:51:10 <aimeeu> yes - no easy way to get to Ethiopian on the MARTA
00:52:01 <ekcs> 7p sounds good too.
00:53:02 <ekcs> ok then are we done with this topic for now?
00:53:06 <thinrichs> Yep
00:53:13 <ekcs> thanks thinrichs !
00:53:17 <ekcs> #topic open discussion
00:54:13 <ekcs> go for it if anyone has anything to discuss =)
00:55:27 <ekcs> who’s going to be there on friday btw? maybe we could do an unofficial outing on friday for those interested. and include some other folks we know/meet too.
00:55:51 <aimeeu> I may be interested (or I may be in AT&T meetings)
00:56:03 <aimeeu> My flight leaves Atlanta at 9 pm
00:56:26 <thinrichs> I'm not there on Friday.  :(
00:56:33 <masahito> I'll be there on Friday.
00:56:58 <ramineni_> I'll be there too
00:57:28 <ekcs> ok got it. maybe i’ll look into that and send out an email.
00:57:56 <ekcs> last three minutes.
00:59:08 <ekcs> just wanna say even though we have found a couple more bugs it seems to be fewer than at the same spot in some of the previous releases in my memory. probably both because we limited features and we did a good job of finding & fixing things early.
00:59:21 <ekcs> good job everyone =)
00:59:56 <thinrichs> Thanks all!
01:00:18 <ekcs> alright then. have a great week all and see you next time!
01:00:34 <ekcs> #endmeeting