23:01:49 <thinrichs> #startmeeting CongressTeamMeeting
23:01:50 <openstack> Meeting started Wed Mar 16 23:01:49 2016 UTC and is due to finish in 60 minutes.  The chair is thinrichs. Information about MeetBot at http://wiki.debian.org/MeetBot.
23:01:51 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
23:01:54 <openstack> The meeting name has been set to 'congressteammeeting'
23:02:23 <thinrichs> ekcs, ramineni, masahito, tsandall: courtesy ping
23:02:30 <tsandall> here
23:02:43 <ekcs> hello
23:03:21 <thinrichs> Seems like a smaller crew today.
23:03:30 <thinrichs> Maybe others are running late.
23:03:36 <thinrichs> Agenda for today…
23:03:45 <thinrichs> 1. Congress-related talks at OpenStack
23:03:58 <thinrichs> 2. PTL nominations are open
23:04:02 <thinrichs> 3. Status updates
23:04:06 <thinrichs> Anything else?
23:05:08 <thinrichs> #topic Congress-related talks
23:05:43 <thinrichs> There are a number of Congress talks happening at the summit…
23:05:51 <thinrichs> Thu 28 11:00a - 11:40a
23:05:51 <thinrichs> Enforcing Application SLAs with Congress and Monasca (fabiog)
23:05:56 <thinrichs> Thu 28 11:50a-12:30
23:05:56 <thinrichs> Congress in NFV-based Mobile Cellular Network Fault Recovery (masahito)
23:06:05 <thinrichs> Thu 28 1:30-2:10p
23:06:05 <thinrichs> Policy Canvas: Draw your policies for OpenStack services (JK)
23:07:34 <thinrichs> #1 is by fabiog, #2 is masahito, #3 is JK who worked with us before
23:08:30 <thinrichs> The 4th talk is on Watcher, which sounds quite similar to Congress but is focused on optimizing resource utilization, instead of policy.
23:08:33 <thinrichs> Thu 28  2:20-3:00p
23:08:33 <thinrichs> Watcher a Resource Manager for OpenStack: Plans for the N-release and beyond (Congress sibling)
23:08:57 <thinrichs> The interesting thing is that these are all scheduled back-to-back.  (Not sure about rooms, but they're all scheduled sequentially.)
23:09:27 <thinrichs> Just figured we'd all want to have those on our calendars so we don't schedule over them.
23:09:31 <thinrichs> Questions or comments?
23:09:49 <thinrichs> BTW you can find all these by searching for Congress on the Austin schedule system.
23:10:22 <ekcs> Sounds great.
23:10:44 <thinrichs> Next topic…
23:10:48 <thinrichs> #topic PTL
23:10:57 <thinrichs> PTL elections are open.
23:11:21 <thinrichs> I'm happy to talk with anyone who might be interested in running.
23:11:58 <thinrichs> I'm running, but if anyone else is interested, please do so.
23:12:12 <thinrichs> It's overall good for the project to rotate PTLs once in a while.
23:12:21 <thinrichs> Or even every cycle.
23:12:31 <thinrichs> (Some projects have decided to have a different PTL on every cycle.)
23:13:31 <thinrichs> Questions or comments?  (I'm happy to take them via email or other means as well.)
23:13:56 <thinrichs> Deadline is soon (don't know when off the top of my head).
23:14:44 <thinrichs> Next topic...
23:14:47 <thinrichs> #topic status updates
23:15:15 <thinrichs> I see no code that needs review before we cut a release.
23:15:34 <thinrichs> Which is a good sign.
23:16:00 <thinrichs> Are there any outstanding bugs that need to be fixed that we know of?
23:16:19 <thinrichs> ekcs: any on your list?
23:16:33 <ekcs> No new bugs from me.
23:16:37 <ekcs> I finally figured out this issue: https://review.openstack.org/#/c/258711/
23:16:38 <patchbot> ekcs: patch 258711 - congress - [WIP] Strange DSE behavior with publish, publish, ...
23:16:38 <ekcs> And determined it's does not affect Congress deployment.
23:16:57 <thinrichs> I saw you made notes.
23:17:11 <thinrichs> So it's a test issue, not a production issue?
23:17:29 <ekcs> Right.
23:17:43 <thinrichs> Phew!
23:18:05 <thinrichs> ekcs: And you tracked down and fixed the bug we discussed last week.
23:18:07 <thinrichs> right?
23:18:19 <ekcs> thinrichs: Right.
23:18:37 <ekcs> Thanks for encouraging further investigation when I initially couldn’t reproduce.
23:19:00 <thinrichs> Sometimes you just get a feeling based on the kind of error.  Thanks for following through!
23:19:10 <thinrichs> Sounds like it's time for me to do a round of testing.
23:19:48 <thinrichs> I'll do that hopefully over the next couple of days and file bugs.  If nothing serious pops up, we'll cut the first release candidate and reopen master for non-mitaka code.
23:19:56 <thinrichs> Sound good?
23:20:02 <ekcs> Great.
23:20:23 <tsandall> sounds good
23:20:42 <thinrichs> ekcs: any other status update to report?
23:21:00 <ekcs> No other reports.
23:21:24 <thinrichs> tsandall: want to give us a status report?
23:21:33 <tsandall> sure
23:22:19 <tsandall> I've been spending most of time reviewing the policy engine implementation and getting a better understanding of the evaluation algorithm, indexing, etc.
23:22:47 <tsandall> I had already reviewed the driver architecture a while ago so I was somewhat familiar with that area
23:23:20 <tsandall> I hope to start reviewing changes as they come in and providing feedback (although it may be a bit low level while I ramp up)
23:23:38 <tsandall> ...and I'm looking forward to Austin.
23:23:43 <tsandall> that's it for now
23:23:47 <thinrichs> Cool!
23:23:52 <ekcs> awesome.
23:24:02 <thinrichs> tsandall: that's definitely the heart of Congress you're looking into
23:24:15 <tsandall> btw, the new DSE code looks great!
23:24:41 <thinrichs> We're definitely excited to get the new DSE code in place!
23:24:58 <clayg> l/wc
23:25:00 <clayg> sorry
23:25:15 <thinrichs> That's it for today from me.
23:25:18 <thinrichs> #topic open discussion
23:25:22 <thinrichs> Anything else to discuss?
23:25:31 <ekcs> Does anyone have a manual testing “script” (non-automated) that they could share?
23:25:31 <ekcs> I'd love to see the kind of cases people are testing and let that inform my manual testing which is rather haphazard at this point.
23:25:32 <ekcs> Not just repeating to exact same tests of course cuz that doesn't do any good.
23:26:35 <thinrichs> The hands-on-lab is probably a good script
23:26:59 <thinrichs> I typically focus on breadth of CLI commands, making sure they work and do what I expect.
23:27:33 <thinrichs> I also try to add a few interesting policy statements, including negation, recursion.
23:27:44 <thinrichs> I also check that all the datasources can connect.
23:28:03 <thinrichs> The CLI is running the API itself, so I don't typically feel the need to run API calls directly.
23:29:14 <thinrichs> ekcs: would you mind testing that the HOL still works (to the extent that makes sense)?
23:29:47 <thinrichs> Not building another instance of the HOL, but rather just running through the exercise with the tip of master.
23:30:02 <ekcs> Ok that helps.
23:30:19 <thinrichs> The other thing we talked about was fixing up the Standalone install instructions
23:30:22 <ekcs> I’ve been trying that but ran into some issues with non-congress stuff. I’ll see if I can resolve and continue.
23:30:46 <thinrichs> We probably don't want to get too hung up on non-congress stuff.
23:30:52 <thinrichs> What's the basic problem?
23:31:37 <ekcs> I got stuck setting up the networks the same way the lab had it at starting state. could be a neutron or horizon issue. but i’m not quite sure yet.
23:31:58 <ekcs> Will revisit. May just adapt the lab exercise to not require the same setup.
23:32:27 <thinrichs> It's hard to get those networks set up properly.  I think we typically do everything from the CLI (not sure if that'll make a difference.)
23:32:36 <thinrichs> Adapting the lab exercise makes sense.
23:32:39 <ekcs> i’ll try that.
23:33:11 <thinrichs> ekcs: maybe I'll try to make a script for the next release
23:33:38 <thinrichs> Anything else?
23:33:47 <thinrichs> bryan_att: I noticed you join earlier.  Anything to discuss today?
23:34:26 <bryan_att> Hi, I am at ONS this week, nothing to discuss - very distracted...
23:34:41 <bryan_att> I'll get back to you next week.
23:34:41 <thinrichs> bryan_att: sounds good.  Enjoy ONS!
23:34:44 <ekcs> Also, this is low-priority, but I figure I'd bring it up because we don't have anything else to discuss anyway.
23:34:45 <ekcs> I ran through the agnostic performance testing code. There is an issue that some of the queries take a long time on first execution but much faster on subsequent executions. The way the tests are set up, running something multiple times and looking at the total time taken can severely underestimate the first-run time. I may patch that up at some point.
23:36:42 <tsandall> just a thought, but the first time queries are run, the indexes are built; subsequent queries would be fast
23:36:56 <thinrichs> tsandall: that makes sense
23:37:15 <thinrichs> ekcs: are you suggesting the tests should be set up so we measure only the first execution?
23:37:20 <thinrichs> That'd be a worst-case kind of analysis.
23:37:41 <thinrichs> Seems valuable to have worst-case as well as more avg-case, which is I guess what we have now.
23:38:37 <thinrichs> Maybe what we want are some tests that do the first-execution only..
23:38:46 <thinrichs> and other tests that do multiple executions.
23:38:47 <ekcs> Not sure exacty. But maybe report the two separately. First execution. And subsequent execution.
23:38:54 <ekcs> right.
23:39:00 <thinrichs> Ideally we'd use these tests so we
23:39:13 <ekcs> but it really comes down to how we interpret the results.
23:39:18 <thinrichs> would know when a commit caused a large performance problem
23:39:43 <thinrichs> Running the perf tests on every commit and rejecting if the test time was much longer than the test without the new code.
23:39:57 <thinrichs> Maybe non-voting
23:41:03 <ekcs> if it’s just difference analysis from commit to commit then it’s not a big issue. But the difference could appear understated. On some of the tests subsequent queries is essentially instantaneous. So if we run it 100 times, whatever difference in first-run performance is / 100.
23:41:40 <thinrichs> Understood.  Makes sense to measure the 2 separately.
23:41:56 <thinrichs> I was mainly just thinking aloud about what we could use these performance tests to do, further down the road.
23:42:11 <thinrichs> Anything else?
23:42:34 <ekcs> When I do devstack.
23:43:29 <ekcs> nova2 never connects. (Not found 404). Haven’t looked much more into it, whether it’s config or whatever. Anyone else see the same?
23:44:10 <thinrichs> I haven't seen that yet, but I also don't remember.
23:45:00 <ekcs> ok.
23:45:27 <thinrichs> I'll take a look when I do my testing pass
23:45:45 <ekcs> Great. nothing else from me.
23:46:06 <thinrichs> Me too.  Let's quit early and spend the time testing.  :)
23:46:19 <tsandall> cya!
23:46:25 <thinrichs> Thanks all!
23:46:27 <thinrichs> #endmeeting