14:02:02 <SergeyLukjanov> #startmeeting sahara
14:02:03 <openstack> Meeting started Thu Jul 16 14:02:02 2015 UTC and is due to finish in 60 minutes.  The chair is SergeyLukjanov. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:02:04 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:02:07 <openstack> The meeting name has been set to 'sahara'
14:02:15 <vgridnev> o/
14:02:34 <weiting> hi
14:02:40 <SergeyLukjanov> #link https://wiki.openstack.org/wiki/Meetings/SaharaAgenda
14:02:42 <sreshetn1ak> o/
14:02:48 <SergeyLukjanov> #topic sahara@horizon status (crobertsrh, NikitaKonovalov)
14:02:52 <SergeyLukjanov> folks, please
14:02:55 <tmckay> SergeyLukjanov, do we have a "talks" topic today?  I have a question for you
14:02:59 <crobertsrh> The move to /contrib has been merged
14:03:02 <tmckay> or maybe open discussion ...
14:03:11 <crobertsrh> Lots of rebasing going on for our existing changes
14:03:14 <SergeyLukjanov> tmckay, do you mean summit talks?
14:03:17 <tmckay> yep
14:03:23 <elmiko> crobertsrh: nice!
14:03:25 <NikitaKonovalov> yes, @crobertsrh but anyway that's a great news
14:03:48 <SergeyLukjanov> tmckay, we could chat in open discussion sections, I've not added the separated topic due to the proposal deadline already happened ;)
14:03:51 <crobertsrh> We shall see how "great" the news is.  Time will tell.
14:04:01 <tmckay> SergeyLukjanov, ok
14:04:09 <SergeyLukjanov> yeah, I hope it'll help us merging patches
14:04:15 <elmiko> +1
14:06:04 <huichun> tmckay:  Trevor ,recurrence spec has been updated as your suggest, thx~ I will do dub migration, so code begin, you can review the spec
14:06:21 <tmckay> huichun, thanks!  I'll look today
14:06:24 <huichun> DB migration
14:06:41 <SergeyLukjanov> #topic News / updates
14:07:31 <NikitaKonovalov> The HDP 2.2 plugin is almost ready. Actually the changes are on review but they are missing some test coverage
14:07:40 <elmiko> working through the first part of the keystone sessions coding, also drafting the api v2 spec, and preparing for castellan integration
14:08:00 <elmiko> i think apiv2 is going to take awhile...
14:08:14 <venza> I started the process of deprecating Spark 1.0.0 from the spark plugin
14:08:36 <kchen> We have already updates the CDH HDFS HA patches. Please help review.
14:08:46 <vgridnev> Generally don't have any updates, I have several internal task, but nt. implementation is finished and ready for your reviews, as same as auto-configs providers
14:09:06 <alazarev> I just returned from parental leave, a lot of code review, thinking how to refactor heat engine code
14:09:39 <NikitaKonovalov> I've started some Rally benchmarking improvements. Those will allow to track hadoop jobs performance.
14:09:45 <tmckay> I need to catch up on reviews, too.
14:10:06 <sreshetn1ak> I'm working on new HDP plugin
14:10:51 <weiting> I am working on the bp for sahara manila hdfs driver.
14:11:16 <sreshetn1ak> all features implemented, I will write spec on next week
14:12:31 <tosky> I pushed the changes for mako templates in test runner, many thanks to degorenko for the support
14:12:39 <tosky> I did some damage on the way, though
14:13:29 <tosky> for example some code needs to be restored because kilo has not the change, and tests from stable/kilo are used when testing stable/kilo
14:14:32 <tmckay> tosky, restored where? in master, or do you mean added to stable/kilo?
14:14:44 <tosky> tmckay: in the sahara-ci-config code
14:15:06 <esikachev> +1
14:15:15 <tosky> there are other issues I see around (for example here https://review.openstack.org/#/c/177280/ ) related to scenario tests not working with kilo
14:15:50 <tosky> there are details to fix (like how to handle some scenario which is valid in master but not in some supported branch - for now just kilo), but SergeyLukjanov promised a spec :)
14:15:58 <vgridnev> tosky, what bad with that?
14:16:47 <tosky> vgridnev: as I mentioned, if we want to always use tests from master, we need to find a way to support all branches with minimal changes
14:17:32 <SergeyLukjanov> tosky, yeah
14:17:43 <SergeyLukjanov> #topic Open discussion
14:17:53 <vgridnev> I believe that we should use different templates for each branch.
14:18:26 <tosky> I would say not different for each branch, but maybe by feature
14:20:39 <vgridnev> Can you give an example about what do you mean in "maybe by feature"? Separate template for feature?
14:20:54 <tosky> for the case above, the template where you specify the constraints should work also with master, so you could call it still vanilla-2.6.0 and call the one with constraints vanilla-2.6.0-something, using it mainly for kilo
14:21:14 <tosky> just thinking aloud
14:21:33 <tosky> this should all be part of the spec
14:22:00 <SergeyLukjanov> tosky, yup
14:22:32 <SergeyLukjanov> tmckay, what's about talks?
14:23:24 <tmckay> oh, thanks
14:24:07 <tmckay> just a logistical note.  The "Lessons learned" talk I see show up twice in my list, both mark "received".  The Sahara/Ironic talk is there once, but not marked "received"
14:24:15 <tmckay> what does all this mean?
14:24:45 <tmckay> Is the Sahara+Ironic talk not submitted correctly, or something?  And why is "Lessons" listed twice?
14:24:56 * tmckay thinks maybe Sergey has been down this road before :)
14:25:49 <SergeyLukjanov> tmckay, hm, looking
14:26:08 <tosky> tmckay: did you receive an email about the talk submission?
14:26:56 <weiting> In my experience, after I submit a topic, it become received, just like you said.
14:27:41 <pino|work> tosky: our is received too
14:27:46 <tosky> yep
14:27:51 <tmckay> I got an initial speaker notification from another talk someone added me to a couple days ago, but I have not got another one since.  My presentation list shows them all, however.  All are marked "received" except for the Sahara+Ironic (but if it wasn't in the system, then I couldn't see it :) ) So I don't know what "received" means
14:28:31 <tosky> for the record, pino|work and me submitted a presentation about libguestfs and image building, which will touch also Sahara images, and an alternative runner (instead of diskimage-builder)
14:28:32 <vgridnev> tosky, I really don't understand what we going to do with scenario tests. So, I would prefer to wait some spec about that. Now I can remove changes for current templates in my patch, is it that ok?
14:28:34 <tmckay> what I mean is, Sergey submitted Sahara+Ironic.  I was added as a speaker
14:29:26 <SergeyLukjanov> tmckay, I've mailed summit@o.o about it, I don't know
14:29:37 <NikitaKonovalov> btw, I also don't see 'recieved" mark on Ironic talk
14:29:38 <tmckay> ok, thanks. Looked fishy to me
14:29:59 <tosky> vgridnev: the direction is to move scenario tests in their own repository and make it branchless like tempest, so that tests from master should work on all supported branches
14:30:05 <tmckay> NikitaKonovalov, ack.  I see you there, but not "received" either
14:30:38 <tosky> vgridnev: this was proposed few months ago, before the Summit, where I think the issue was discussed (but I was not there), and SergeyLukjanov will send a spec about this
14:30:50 <tmckay> SergeyLukjanov, did you also mention to them that "Lessons learned" is listed twice?
14:30:57 <tmckay> that seems equally bad
14:31:02 <tosky> vgridnev: that said, the details on how to implement this, like how to deal with cases like this, are not clear (at least for me) yet
14:31:53 <SergeyLukjanov> tmckay, will do
14:32:00 <vgridnev> So, anyway I will remove all changes until full design is not approved
14:32:17 <vgridnev> changes for templates*
14:32:18 <tosky> vgridnev: if the existing scenario works even after your changes (which should be, you are "simply" not tracking the new feature), I would say: either create a new template with a different name (better), or remove it for now
14:32:24 <elmiko> speaking of all this, NikitaKonovalov did you get the invite to the sparkhara talk?
14:32:55 <tosky> if anyone else have some other idea or suggestion about this (templates and branchless tests, please share!
14:33:25 <NikitaKonovalov> elmiko: I see myself added to the talk, but no email if you mean that
14:33:34 <elmiko> bummer...
14:34:28 <NikitaKonovalov> and it is marked recieved, whatever it means
14:35:00 <elmiko> yea, i saw it was recieved, just wanted to make sure you got the email. but i imagine we'll have to wait or something
14:35:35 <NikitaKonovalov> elmiko: what the email topic should look like? may be I've missed it
14:36:04 <elmiko> NikitaKonovalov: not sure what the subject will be
14:36:26 <elmiko> oh well, you're on there. that's good enough =)
14:37:33 <weiting> The subject title should be like "OpenStack Summit - Tokyo 2015 - Speaker Notification"
14:38:43 <elmiko> thanks weiting
14:43:07 <tmckay> I think it might only send a single "Speaker Notification".
14:43:23 <tmckay> If you are listed on multiple talks, I am not sure you will get another after the first one
14:43:57 <elmiko> ah, ok
14:46:11 <SergeyLukjanov> no ideas about the submissions, hope, folks will answer
14:46:20 <SergeyLukjanov> there were no issues last two years ;)
14:46:30 <SergeyLukjanov> anything else to chat about today?
14:47:48 <elmiko> nothing from me
14:48:27 <tmckay> SergeyLukjanov, problem with scale :)  2 years ago it was smaller
14:48:36 <tmckay> nothing else from me
14:49:24 <tosky> nothing else from me
14:50:13 <weiting_> I got one question,
14:50:28 <weiting_> Have you ever take a look at Amazon Data Pipeline
14:50:51 <weiting_> A new solution from Amazon
14:50:56 <elmiko> interesting
14:51:02 <weiting_> I think it is a good idea for Sahara EDP
14:51:22 <elmiko> do they have a public api?
14:51:49 <weiting_> Yes, it also have include a rest api
14:51:59 <elmiko> neat
14:53:42 <alazarev> weiting_, just read description, can't get the benefits, one more abstract layer?
14:53:52 <weiting_> Anyway, I just thought this is what the customer want, they would like to have a way to control the data flow.
14:54:10 <weiting_> alazarev, yes another abstract layer
14:54:38 <alazarev> weiting_, we have it too, called "datasource" :)
14:55:10 <weiting_> alazarev, yes, a poor version I think.
14:55:23 <weiting_> :(
14:56:12 <huichun> I just think Ozzie engine may be not stable for edp jobs, performance issue
14:56:32 <huichun> maybe we can find a substitution
14:57:47 <SergeyLukjanov> 2 mins left
14:58:28 <alazarev> huichun, console calls? like for spark
14:58:32 <SergeyLukjanov> #endmeeting