18:00:12 #startmeeting sahara 18:00:13 Meeting started Thu Jun 23 18:00:12 2016 UTC and is due to finish in 60 minutes. The chair is vgridnev. Information about MeetBot at http://wiki.debian.org/MeetBot. 18:00:14 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 18:00:17 The meeting name has been set to 'sahara' 18:00:19 hello folks! 18:00:23 hey 18:01:42 hi ! 18:01:42 hi 18:02:08 o/ 18:03:24 #link https://wiki.openstack.org/wiki/Meetings/SaharaAgenda 18:03:27 o/ 18:03:36 #topic News / updates 18:04:09 my status shortly: https://review.openstack.org/#/q/status:open+project:openstack/sahara+branch:master+topic:bp/plugin-management-api 18:04:49 i put the spec for storm python jobs up for review yesterday https://review.openstack.org/#/c/333003/, create bp for that and storm upgrade on which i'm currently working on 18:05:10 I helped fixing jobs on the main CI 18:05:21 o/ 18:05:29 hey 18:05:36 Working through some licensing issues around python-libguestfs bindings (not currently in pypi.) Hopefully those will be resolved soon and I can push the change. 18:05:57 sorry, i am preparing for a conference talk/demo, my bandwidth is really shot to hell this week and next 18:06:03 working on publisher job for devstack logs from sahara-ci labs 18:06:41 (The bindings are perfectly stable and mature, but the authors had some concerns about Python Foundation legal language. Working on it; will hopefully have an elegant solution soon, and if not, I'll post a hackier solution.) 18:06:50 only sahara related things i've done this past week was a blog post, https://elmiko.github.io/2016/06/22/hardening-sahara.html 18:08:27 hello/ 18:08:38 hello crobertsrh 18:09:24 crobertsrh: Nice merge of hello and wave emote. Raising the bar for IRC greetings right there. 18:09:59 I try :) 18:10:34 other news updates? 18:10:48 I've just been trying to do some reviews (and failing mostly) 18:10:56 I need to do some spec reviewing 18:11:37 #topic Review priorities 18:11:41 #link https://etherpad.openstack.org/p/sahara-review-priorities 18:12:18 we were discussing voiding additional requirements for docs building in saharaclient previously 18:12:26 #link https://review.openstack.org/#/c/326582/ 18:13:10 news: seems like this is *not* going to be landed to globals, since lib is not maintained. 18:13:23 #link https://review.openstack.org/#/c/325932/ 18:13:36 Well, if the lib isn't going to be maintained, maybe it's actually *better* for us to copy what we're using. 18:13:54 as author of lib answered, he is searching for maintainer 18:14:04 So... I guess that's a dramatic reversal of my previous position. :) 18:14:52 copy of his answer, http://paste.openstack.org/show/521722/ 18:16:34 Yeah. Sounds like this patch is our best solution. 18:16:47 (Given that what the modules are doing for us is really quite nice.) 18:17:12 Thanks for making sure it was the right answer. 18:17:35 +1, seems like "it's our problem now" 18:18:45 ok, nice 18:18:46 Hopefully it'll just sit there cranking out docs and we'll never think about it again. 18:19:30 #topic API v2 progress 18:19:39 is there some news? I bet no 18:20:14 except of some changes of mine https://review.openstack.org/#/q/status:open+project:openstack/sahara+branch:master+topic:bp/plugin-management-api 18:20:37 elmiko, ? 18:21:10 oop, sorry 18:21:13 nothing new to report 18:21:40 #topic Specs / features status 18:22:06 #topic Initial kerberos integration 18:22:17 #link https://review.openstack.org/#/c/327616/ 18:22:28 any other questions to resolve? 18:23:11 vgridnev: Looks ok to me. 18:23:14 i think it's looks good 18:24:23 do we need to wait some other votes? 18:24:39 the last version i looked at was good to me 18:25:07 elmiko: I think if we get your +2 we can merge; it's been open for comment for awhile. 18:26:22 ok, moving 18:26:24 spec looks good to me as well 18:26:47 I don't why not move on with this either 18:26:51 (Or someone else's :) ) 18:27:00 egafford: yea, i will read it now 18:28:14 NikitaKonovalov, I guess egafford hinted that you have +2 vote 18:28:42 I did hint that! 18:29:19 #topic Allow creation of python topologies for Storm 18:29:26 #link https://review.openstack.org/#/c/333003/ 18:29:59 The spec looks good; I asked for a little bit more detail on what the cmdline looks like and whether any special configs are needed (and how data sources get passed in.) 18:30:01 tellesnobrega, it's your time to address comments on spec 18:30:05 so that, anyone had a chance to read it yet? (except egafford - saw your comments and will improve it) 18:30:09 Nothing wrong with it at all though. 18:30:23 thanks vgridnev 18:30:49 basically I'm proposing that we allow in storm to be submitted jobs implemented using the pyleus framework 18:31:18 which allows the user to develop the whole thing using python and yaml (which is quicker and friendlier than java) 18:31:47 this won't change much of the way storm works today, just a single command line change depending on the job type select (storm or pyleus) 18:32:20 I will address Elise's comments on the spec but if anyone wants to take a look on that today I can wait to do it all at once 18:32:48 vgridnev, I don't think UI changes will be needed 18:33:48 tellesnobrega: I believe that at present any new job type requires a UI change in order for it to be useable via the UI. crobertsrh, NikitaKonovalov, is this still true? 18:34:02 Most likely, yes 18:34:46 tellesnobrega, so just noted, what is the scope of using pyleus (cluster nodes / controller with sahara) ? I mean, do we need to push that to requirements? 18:34:47 The job types dropdown did get smarter awhile back, but the UI might need tweaks for the specifics of the job 18:34:51 egafford, I see, you mean adding it to the list of job types and so on... yes, that will be needed 18:35:13 tellesnobrega: Yeah. Should be really minor, but. 18:35:35 vgridnev, cluster nodes will need to have it installed, sahara controller won't 18:35:55 cool 18:36:28 the best way to do that is to install on image generation? 18:36:59 tellesnobrega: changes to UI mostly depend on how much stuff the job type supports. If it woks with Data Sources, then you'll need to enable them in UI. But that again is pretty minor. 18:37:33 NikitaKonovalov, makes sense, noted :) 18:37:41 tellesnobrega: Yeah, hopefully you can just install the packages via SIE. 18:37:50 egafford, awesome :) 18:38:03 (Assuming pyleus is packaged for our distros.) 18:38:20 egafford, I have to check that out 18:38:29 (Otherwise your life becomes slightly more complex but not very much more complex.) 18:38:35 is it a native python package? 18:38:52 if it is, it shouldn't be difficult to ask distros to package it 18:38:52 yes, installed by pip 18:39:09 I don't think this will be a problem 18:39:25 good 18:39:34 but it is good to have it in mind 18:39:53 +1 to both not a problem and good to have in mind. :) 18:40:13 :) 18:41:56 ok, tellesnobrega could you please update spec then with notes from discussion? 18:42:29 vgridnev, will do 18:42:35 ok, cool 18:42:48 #topic Open discussion 18:43:22 kerberos spec lgtm, shall i merge it? 18:43:34 elmiko, sure 18:44:01 ok, approved =) 18:44:15 \o/ 18:44:21 elmiko, this one also can be merged based on our discussions today https://review.openstack.org/#/c/326582/ 18:44:24 a really nice step forward for security 18:44:41 I like this whole "let's do specific things during our meetings" approach. 18:45:02 yea, it can work really well 18:45:10 vgridnev: ack, looking at it now 18:46:11 elmiko: In case you missed it earlier, basically there's no one maintaining that lib, it's legal for us to copy it, and it's doing really nice stuff for us in terms of doc automation, so code copy is actually the right answer. 18:46:18 (Seldom is, totally is here.) 18:46:44 egafford: yeah, saw the notes on the review. as long as we are good taking on the maintenance of that for our purposes, i'm good 18:47:30 added my +2, i'll let egafford do the +a since my familiarity is low on this one ;) 18:48:05 BOOM. WF. Basically I'm for maintaining it as part of the codebase until it becomes unworth it. 18:48:34 (And I doubt it will; it seems quite likely that it'll continue working for the forseeable future as-is.) 18:50:03 yup, i'm good with that, just wanted to raise the question =) 18:50:36 there where an question in IRC 18:50:43 btw, I am preparing my trip to NYC for the Operatos Mid-Cycle. Is anyone from the sahara project going there ? 18:51:13 just to make sure, is anyone from RedHatters going to attend? 18:51:22 not me 18:51:25 Wasn't going to, though I'm almost certainly the closest human to it. 18:51:38 yeah, you could just take the train down there 18:51:51 * egafford googles the schedule... 18:52:16 vgridnev: Was there a specific request for someone at it? Like, does someone need help with an evaluation or somesuch? 18:52:29 Or was it more general? 18:53:33 egafford, actually it's was an operator who was asking for sahara question about sahara. As far as I know, he is also going to future summit 18:53:58 sahara question about sahara, lol 18:54:27 Those are the best kind of Sahara questions, and the best kind of questions about Sahara. :) 18:54:41 Ok, if he's going to summit we can talk to him then. 18:55:15 (Without additional travel budget.) It's zioproto on freenode who had questions? 18:55:29 yes 18:55:57 http://lists.openstack.org/pipermail/openstack-operators/2016-June/010674.html 18:56:41 yep, he is using our hadoop swift library for accessing data 18:57:27 ok, seems like there are no other points to discuss 18:57:36 #endmeeting