17:00:17 <hogepodge> #startmeeting refstack 17:00:18 <openstack> Meeting started Tue Jan 9 17:00:17 2018 UTC and is due to finish in 60 minutes. The chair is hogepodge. Information about MeetBot at http://wiki.debian.org/MeetBot. 17:00:19 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 17:00:22 <openstack> The meeting name has been set to 'refstack' 17:00:33 <hogepodge> #link https://etherpad.openstack.org/p/refstack-meeting-18-01-09 agenda 17:00:35 <tosky> o/ 17:00:59 <mguiney> o/ 17:02:18 <hogepodge> Hello everybody. 17:02:27 <hogepodge> Please update the agenda if you have any items 17:02:58 <hogepodge> luzC_: is still out, set to return on January 15 17:03:47 <hogepodge> #topic PTG 17:05:05 <hogepodge> A reminder that we're around six weeks away from the next PTG. We'll have space, and if there's something you really want to collaborate in person on there get it on the agenda. As we get closer I'll start putting together 40 minute sessions devoted to individual topics. 17:06:07 <hogepodge> #topic RefStack 17:06:27 <hogepodge> This review is passing the gate 17:06:27 <hogepodge> https://review.openstack.org/#/c/507695/ Disable Anonymous Upload 17:06:30 <hogepodge> #link https://review.openstack.org/#/c/507695/ Disable Anonymous Upload 17:06:43 <mguiney> \o/ 17:07:29 <hogepodge> I'd like to call it out., particularly lines 326 and 345 here https://review.openstack.org/#/c/507695/23/refstack/db/sqlalchemy/api.py 17:09:21 * mguiney nods 17:09:29 <hogepodge> Python2 and Python3 handle strings and bytes in totally different ways, and to get the gate to pass on both versions I had to drop the 'encode' call. I don't know if it's strictly necessary, or if there's a way to choose between Python2 and Python3 implementations. I'm a bit worried that I broke something, so if folks more versed in the 2->3 transition and check my work that would be greatly 17:09:29 <hogepodge> appreciated. 17:10:23 <hogepodge> But, I do know that as it stood RefStack did not work with Python3. That code path would have always failed. 17:10:34 <clarkb> encode ascii there makes a string objcet with ascii characters in python3, which is not a bytestring 17:11:07 <clarkb> based on a quick read that is likely safe since ascii is 7 bit subset of 8bit bytestrings 17:11:32 <hogepodge> clarkb: on key.encode('ascii'), key is a bytes object 17:11:45 <clarkb> hogepodge: ya and the encode converts it to a string object in python3 17:11:56 <clarkb> which base64 doesn't like 17:12:02 <clarkb> (so I think the change is safe as is) 17:12:04 <hogepodge> in Py2 is does something, and in Py3 is throws 'bytes does not have a method encode' 17:12:27 <hogepodge> I don't know why I keep typing "is" instead of "it". :-/ 17:12:32 <hogepodge> clarkb: cool :-) 17:12:37 <clarkb> oh right needs to be decode? in any case I think its safe because base64 wants bytes anyways so no conversion necessary 17:13:20 <hogepodge> that's what I was thinking, but I was a bit worried that we were tuning the tests to pass and not checking the actual behavior 17:14:30 <hogepodge> Since this is a standalone application, I'd like us to move towards dropping Py2 support and going all-in on Py3. 17:14:52 <hogepodge> At least on the server side. Opinions or thoughts on that? 17:15:55 <mguiney> i've mentioned it previously, but all for it 17:18:06 <hogepodge> Speaking of updates. pvaneck has started some node4 -> node6 transitions 17:18:16 <hogepodge> #link https://review.openstack.org/#/c/528831/ Update Node from v4 vo v6 17:19:29 <hogepodge> On my development builds of RefStack, I'm also seeing a bunch of deprecation warnings on the NPM libraries we're using. We should get ahead of future breakages (application, security, and otherwise) and start migrating to never versions of the libraries. 17:19:44 <hogepodge> s/never/newer/ 17:21:51 <hogepodge> Quiet meeting today :-) 17:22:05 <hogepodge> we can revisit these topics next week once luzC_ is back 17:22:07 <tosky> NPN transition, brrr 17:22:38 <hogepodge> I know almost nothing about Javascript, so it will be a learning experience indeed. 17:23:13 <hogepodge> Other upcoming RefStack work that we're hoping to move to the front burners 17:23:29 <hogepodge> #link https://review.openstack.org/#/c/530681/ Subunit Upload API 17:23:40 <hogepodge> #link https://review.openstack.org/#/c/528844/ New Doc PTI Jobs 17:24:08 <mguiney> subunit upload api will have a new patch out (hopefully) today 17:24:21 <hogepodge> The Docs PTI jobs lgtm 17:24:24 <hogepodge> mguiney: cool! 17:25:39 <hogepodge> looking forward to it 17:25:51 <mguiney> have been chipping away at it, apologies for any delay 17:26:03 <hogepodge> Only so many hours in the day. :-) 17:26:20 <hogepodge> We have a pretty big plate of RefStack work for such a small group. 17:26:45 <hogepodge> No current work on the refstack-client 17:27:23 <mguiney> would it be worth adding a keysign upload functionality on that front, ooc? 17:27:38 <mguiney> given that you literally have to log in via the gui for the current setup to work 17:27:51 <hogepodge> yes 17:27:57 <mguiney> and the sign is already generated via refstack-client already 17:28:00 <hogepodge> would we need an endpoint to do that? 17:28:11 <mguiney> i think... maybe, yes 17:28:17 <mguiney> but only with a post 17:28:33 <mguiney> because it just adds more entries for your user 17:28:34 <hogepodge> I wonder if we don't allow it because you have to bootstrap your key somehow. 17:28:45 <mguiney> what do you mean? 17:28:57 <hogepodge> You have to establish your identity to associate your public key with your account. 17:29:15 <hogepodge> On the web interface you do that with a password and a session. 17:29:23 <mguiney> true. hm. 17:29:43 <hogepodge> It's probably a good topic to sort out during the PTG. 17:29:49 * mguiney nods 17:29:51 <mguiney> cool 17:30:07 <mguiney> just drives me up the wall, not much a fan of having to gui :) 17:30:18 <hogepodge> of course :-) 17:30:43 <mguiney> cool 17:30:50 <hogepodge> I feel like in every application I use that needs a public key I upload it through a web interface. Gerrit and GitHub stand out in my mind 17:31:45 <hogepodge> Ok, moving on. 17:31:57 <hogepodge> #topic Python TempestConf 17:32:06 <hogepodge> tosky: you want to lead this part of the meeting? 17:34:05 <tosky> so, chandankumar is writing some documentation on the proposed changes for tempestconf 17:34:14 <hogepodge> #chair tosky 17:34:15 <openstack> Current chairs: hogepodge tosky 17:34:21 <tosky> the thing that we discussed briefly before the holidays and last week 17:34:34 <tosky> basically collecting how we generate each key 17:34:49 <tosky> let me share a link... 17:35:05 <tosky> #info (draft) https://etherpad.openstack.org/p/tempestconf_improvements 17:36:17 <tosky> while waiting on this big work, we just briefly discussed whether to start releasing a new version for Queens, and yeah, probably we need one for the keystone v3 changes 17:37:03 <tosky> if there are no questions, let's go with the open reviews 17:37:10 * tosky waits few seconds 17:37:23 <hogepodge> lgtm 17:38:03 <tosky> #link https://review.openstack.org/#/c/513330/ Add initial .zuul file, devstack and packstack job 17:38:31 <tosky> the last revision finally works! But it require a change in packstack 17:38:50 <tosky> the reviewers asked for a small documentation change which is now applied, so I guess we will start nagging them soon 17:39:26 <tosky> when that packstack change is merged, we can merge our, which introduces the testing with devstack and packstack, and both with admin and demo user 17:39:34 <hogepodge> what was the packstack change? 17:40:09 <tosky> it's the Depends-On of the review: https://review.openstack.org/#/c/520594/ 17:40:53 <tosky> basically a new packstack "profile" which does not run tempest (in our tests tempestconf obviously takes care of that) 17:41:19 <tosky> and moving most of the code to a role, so we can reuse it 17:41:29 <tosky> niceness of zuulv3 17:42:14 <tosky> if there are no other questions on this... 17:42:52 <tosky> #link https://review.openstack.org/#/c/519714/ Add EC2-API support 17:43:19 <tosky> it is still a work in progress (the author set -W) 17:43:24 <tosky> so no much to discuss yet 17:43:35 <tosky> and the last one is: 17:43:38 <tosky> #link https://review.openstack.org/#/c/528380/ Make tempest setup image verification aware 17:44:50 <tosky> the RDO packaging was fixed adding the new dependencies required by the patch (unfortunately it's not possible to patch the RPM spec on the fly just for a patch) 17:45:10 <tosky> and Arx reviewed it, so waiting for the answer from the author 17:45:48 <tosky> the idea seems fine, the code should be just be tuned to avoid dumping even more unstructured code in the main python file while we try to restructure it 17:45:57 <tosky> and that's it from my side 17:46:59 <hogepodge> heh, ok 17:48:48 <hogepodge> I'll wait for some of the patches to move out of wip 17:49:17 <chandankumar> ah, Sorry i missed the meeting 17:50:33 <hogepodge> hi chandankumar! :-) 17:50:45 <hogepodge> tosky: anything else on this topic? 17:51:13 <tosky> nothing else 17:51:22 <hogepodge> #topic open discussion 17:51:28 <hogepodge> Anything else anyone wanted to chat about? 17:53:25 <hogepodge> Ok, everyone gets six minutes of their day back. 17:53:37 <hogepodge> #endmeeting