17:01:14 #startmeeting refstack 17:01:15 Meeting started Tue Jan 16 17:01:14 2018 UTC and is due to finish in 60 minutes. The chair is hogepodge. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:01:16 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:01:18 The meeting name has been set to 'refstack' 17:01:34 #link https://etherpad.openstack.org/p/refstack-meeting-18-01-16 agenda 17:01:58 Hello everybody! Please add any additional items to the agenda. I'll give folks a few minutes to arrive. 17:03:34 o/ 17:03:37 o/ 17:04:27 ping luzC_ 17:04:59 ok, small crew today, but let's get started 17:05:15 #topic upcoming meetings 17:05:48 I will be at our staff offsite next week, so won't be available. We can either run the meeting or cancel it. 17:06:04 (I would have mguiney host the meeting if we wanted to have it) 17:06:57 With the light load, I'm inclined for us to just cancel it. 17:07:23 i may send out an email of pending patches/patches up for review, if we do 17:08:21 does anyone want to have the meeting next week? 17:09:04 tough crowd ;-) 17:09:29 i mean, i would, but as one of the people with patches up/incoming, i'm a bit biased 17:10:08 mguiney: sounds great to me. I'll have you run the meeting and if I can drop in I will. 17:10:15 you can always ping people around on the channel; no strong opinion on my side 17:10:19 * mguiney nods 17:10:22 ok 17:11:24 #topic PTG 17:11:47 #link https://etherpad.openstack.org/p/InteropDublin2018PTG Planninc Doc 17:12:03 Reminder of the upcoming PTG. 17:12:22 We'll be splitting time between InteropWG and RefStack, so drop your topics in. 17:13:09 A few topics I'll want to cover will be dropping Python 2 support, and moving RefStack into more of a maintenance phase after some key features land. 17:13:45 We'll also have space to talk about python-tempestconf and how it relates to the refstack-client 17:13:53 dropping Python 2 for all projects, or "just" the RefStack server? 17:14:05 Just the RefStack server. 17:14:21 ah, ok - because if it applies to the client too, we should accelerate the py3 support for tempestconf too 17:14:29 I mean, we should implement it anyway, but... 17:15:31 In the last patch we landed, we uncovered some code that wasn't py3 compatible, and I personally feel that supporting two code-bases on a server with a very small deployment footprint doesn't warrant trying to support two versions. I'd rather chase down all the python2'isms and make it py3 clean. 17:16:00 yes, for the server it definitely makes sense 17:16:08 On the user facing side, we have so many users that are on systems like CentOS 7 and OS X (where py27 is default) that I don't want to make life harder for them. 17:17:52 But, returning to the topic, please fill out the etherpad with your desired work items, and I'll start putting together a schedule (which we can adjust to account for conflicts with other project work) 17:18:17 * mguiney nods 17:19:11 #topic PTL Elections 17:19:22 uh, already that time 17:19:37 It's PTL election season. Anyone want to take leadership of the team? 17:20:48 I'm happy to keep managing it, but my vision is to get us to a more stable place and put us into official maintenance mode (on the server side at least). If someone has grander ambitions this would be your chance to step in. 17:21:18 python-tempestconf would continue to be actively developed, because it's a young project and of broader interest to the community. 17:21:49 well, the maintaince mode can always be revisited if new requirements/code/workload/TC decisions come 17:21:59 * mguiney nods 17:22:33 :-) 17:23:14 We still have features we need to develop, so it's not happening right now. But by mid year I'd like the server to just be chugging along and doing it's work 17:23:15 i'd be glad to run, but am also still new enough that i dont want to step on the toes of someone more experienced 17:24:30 The additional responsibilities are mostly in organizing and running the meetings. 17:24:59 * mguiney nods 17:25:08 We can chat about it later today :-) 17:25:08 i can definitely do that then 17:25:12 cool 17:25:37 #topic current reviews 17:25:44 Disable Anonymous Upload 17:26:05 merged, finally! 17:26:16 Patch merged after one bit of testing (who would have guessed that u"False" evaluates to True?) Thanks mguiney for catching that. 17:26:44 Work items to finish off that task are here 17:27:09 #link https://etherpad.openstack.org/p/refstack-disable-anonymous-upload 17:27:21 (oh wait, that etherpad is empty... I thought we had a list of work items) 17:27:42 I'll dig through our meeting history and paste them into that therpad. 17:28:17 Need to make the announcement that we're disabling uploads, make the switch on the server, then run mguiney's script for marking anonymous data that was used for validation. 17:28:29 was probably in some similarly named pad 17:28:59 copy-paste left us with a dangling reference. 17:29:01 should we make one more check that the current dataset is up to date, as well? 17:29:50 mguiney: yes, new test results have come in. 17:30:15 We'll have to create a new dataset based once anonymous uploads have been disabled. 17:30:49 #link https://review.openstack.org/#/c/528831/ npm4 -> npm6 documentation 17:31:22 pvaneck has a patch to update documentation from npm4 -> npm6 17:31:55 but he rightly states we need to switch the gate jobs over to npm6 to make sure the server actually works with it. So this is one of the work items we'll have to take 17:32:07 #action update gate jobs to use npm6 17:32:24 either mguiney or I could take that 17:33:18 We can work it out this afternoon during our 1:1 17:33:28 can do! 17:34:22 #link https://review.openstack.org/#/c/530681/ Subunit Upload API 17:34:51 Major feature for RefStack. Hoping we'll be able to seriously turn our attentions on implementing this one. 17:35:14 this is coming along, though i am currently working with a small issue with getting it to use the proper conf 17:35:47 as soon as i can get that resolved, i should have a new chunk of code to upload 17:36:25 ok, we should check in with mtreinish to see if he has any ideas on how to smoothly integrate the tooling 17:36:35 can do 17:37:15 RefStack API documentation is deferred until the higher priority work has merged. 17:37:32 And with that I'm going to turn it over to tosky for python-tempestconfig work 17:37:39 #topic Python TempestConf 17:37:46 #chair tosky 17:37:47 Current chairs: hogepodge tosky 17:38:28 so, first a note about a review *not* in the list: packstack and devstack tests are now running for each change 17:38:43 the review was finally merged - with some extra kicking of zuul 17:38:52 and now, about the open reviews... 17:39:14 #link https://review.openstack.org/#/c/528380/ Make tempest setup image verification aware 17:39:44 untouched since last week, I guess we need to ping the author again; there are some comment from arxcruz|off 17:40:28 I'm going fast with the list, we can go back to any of them in case of questions 17:40:49 #link https://review.openstack.org/#/c/519714/ Add EC2-API support 17:41:01 the author removed +w, but there is a small final change requested 17:41:08 I think/hope that it will be merged soon 17:41:56 now two changes related to refactoring, I haven't checked them yet; the behavior should be unchanged, but they should make the evolution of the code easier 17:42:09 #link https://review.openstack.org/#/c/533138/ Split main to smaller functions 17:42:15 (the title says it all) 17:42:42 #link https://review.openstack.org/#/c/533434/ Split ClientManager 17:42:52 moves one class and related code to a different file 17:43:13 I suspect that we may want to release a version before going with refactoring, even if limited 17:43:36 chandankumar, who could not attend today, agreed that we could have a release now, around Queens-3 17:44:00 and the final review from the list, just landed, a small but useful optimization 17:44:04 #link https://review.openstack.org/#/c/533777/ Don't refetch same image over and over 17:44:20 zuul timed out, but maybe it's related to the gates - let's see 17:44:36 and that's all from my side for today, unless there are questions 17:44:59 Yes, zuul is getting spectre/meltdown patches, and is being interrupted 17:45:07 jobs in progress need to be rechecked 17:45:34 last updates landed this morning, so the gate should be stable going forward 17:45:43 there were different queue restarts in the last days 17:45:49 hopefully stable :) 17:45:57 we are also upgrading to a bigger server 17:46:07 I just need to get my hands on the code to review 17:46:12 hi clarkb! :-) 17:47:34 yep, I used the shorter queue during Saturday to merge a review for tempest (and Martin Kopec did the same for the python-tempestconf review that added the test jobs :) 17:48:28 #topic open discussion 17:48:35 Anything else anyone wanted to chat about? 17:51:00 ok, have a great Tuesday! 17:51:04 #endmeeting