18:01:27 <SergeyLukjanov> #startmeeting sahara
18:01:27 <openstack> Meeting started Thu Jun 26 18:01:27 2014 UTC and is due to finish in 60 minutes.  The chair is SergeyLukjanov. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:01:28 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
18:01:30 <openstack> The meeting name has been set to 'sahara'
18:02:09 <SergeyLukjanov> #link https://wiki.openstack.org/wiki/Meetings/SaharaAgenda
18:02:18 <SergeyLukjanov> #topic News / updates
18:02:20 <SergeyLukjanov> folks, please
18:02:59 <SergeyLukjanov> #info today was the PTL Update Webinar and I've presented some info about Icehouse results and Juno plans
18:02:59 <elmiko> i've been putting more work in on the swift/auth issue. i've got an updated plan on the private bug report and i've been doing some local testing of the trust/delegation mechanisms.
18:03:27 <crobertsrh> Merge is still an ongoing process.  Finally getting some core reviewers in on the action.  I have a standing agenda item in their weekly meeting.
18:03:44 <SergeyLukjanov> elmiko, awesome, it looks very good
18:04:05 <alazarev> PTO, bugs
18:04:06 <sreshetnyak> o/
18:04:10 <elmiko> SergeyLukjanov: i'm working on a local copy of a spec for when we are ready
18:04:31 <alazarev> investigaton about Secured Hadoop
18:04:40 <SergeyLukjanov> crobertsrh, cool, please, let me know if we'll have an issues again
18:04:56 <sreshetnyak> i'm working on hadoop 2 refactoring for support hadoop 2.4.0
18:05:04 <SergeyLukjanov> crobertsrh, I have an action item for openstack project meeting each other week
18:05:18 <crobertsrh> will do.  Most of the recent comments have had to do with translation blocks, so nothing really significant being found.
18:05:43 <aignatov> no special from me :( was busy on other internal stuff
18:05:45 <SergeyLukjanov> crobertsrh, I hope that someday I'll wake up and see everything merged ;)
18:06:01 <elmiko> i share your hope
18:06:08 <SergeyLukjanov> oh, I have a notice
18:06:09 <crobertsrh> For me, it is merged in my dreams, but when I wake up, my dreams are shattered :(
18:06:11 <dmitryme> aignatov: me too :-(
18:06:14 <aignatov> plan to update and move samples from etc to public docs
18:06:19 <SergeyLukjanov> I'll be traveling a lot next 1.5 weeks
18:06:29 <aignatov> SergeyLukjanov: do I need to create spec for it?
18:06:53 <SergeyLukjanov> aignatov, you can do as one more example
18:07:20 <SergeyLukjanov> currently we have one merged spec from Trevo
18:07:23 <SergeyLukjanov> Trevor*
18:07:25 <aignatov> so I should do, right? :)
18:07:30 <SergeyLukjanov> #link https://github.com/openstack/sahara-specs/blob/master/specs/juno/edp-move-examples.rst
18:07:34 <SergeyLukjanov> aignatov, yup
18:07:53 <aignatov> I saw it, thx, nice example :)
18:08:02 <SergeyLukjanov> any other updates and/or comments?
18:08:20 <alazarev> also emailed to openstack-dev about hadoop components upgrade, but this is separate topic
18:08:32 <SergeyLukjanov> alazarev, yeah
18:08:42 <SergeyLukjanov> #topic Action items from the last meeting
18:09:03 <SergeyLukjanov> #action aignatov create bp re moving/updating rest samples docs and do it
18:09:13 <SergeyLukjanov> #info DONE: alazarev reraise discussions about backward compat https://review.openstack.org/#/c/98260/
18:09:25 <SergeyLukjanov> #action SergeyLukjanov to create bp with steps to enable heat be default
18:09:37 <SergeyLukjanov> #action SergeyLukjanov to create bp about removing/hiding username@image for heat based provisioning
18:09:53 <aignatov> SergeyLukjanov: move my AI to the next again :)
18:10:01 <SergeyLukjanov> aignatov, already done ;)
18:10:22 <aignatov> ah, I see, didn’t realised that you was creating it :)
18:10:49 <SergeyLukjanov> #topic Review of backward compat
18:10:55 <SergeyLukjanov> #link https://review.openstack.org/#/c/98260/
18:11:46 <SergeyLukjanov> #link http://lists.openstack.org/pipermail/openstack-dev/2014-June/038565.html
18:11:46 <crobertsrh> I should get back on track with sahara reviews....too much time in horizon land for me
18:11:58 <alazarev> actually I’m Ok with solution to make it vendor specific
18:12:26 <SergeyLukjanov> crobertsrh, it'll be really great if you have some review power after horizon :)
18:12:35 <alazarev> if user can upgrade ambari agent in existing cluster there will be no problems with scaling
18:12:40 <crobertsrh> Yes
18:12:44 <alazarev> just need to document upgrade procedure
18:12:50 <bob_nettleton> alazarev, ok, this is good to hear.  is there anything else holding up this merge?
18:13:29 <SergeyLukjanov> alazarev, but the version in sahara db will be still old and sahara will not be able to do anything with this cluster
18:13:38 <bob_nettleton> alazarev, the other option is a manual de-scaling for some compatibility scenarios, as Erik mentioned in the email yesterday.
18:14:18 <SergeyLukjanov> and one more concern was that if we change some plugin internals in a way that it changes deployment layout than it could be broken too
18:14:40 <SergeyLukjanov> but generally I see that we have no good solutions for all of the concerns listed on this topic
18:15:03 <SergeyLukjanov> so, I'm not really against merging upgrade for hdp
18:15:30 <bob_nettleton> SergeyLukjanov, great.  thanks!
18:16:29 <aignatov> I’m ok too with that too, Ambari could be upgraded, actually I put my +1 on this patch :)
18:16:50 <SergeyLukjanov> aignatov, but version in sahara db will be still old one
18:17:11 <ErikB> What is the version is Sahara DB indicate?
18:17:25 <ErikB> sorry, what does the version in Sahara DB indicate?
18:17:45 <bob_nettleton> SergeyLukjanov, what version are we talking about in the sahara db?  I'm confused, because we still support the same versions of HDP, we are just moving up to using a newer Ambari to handle the implementation details of the deployment.
18:18:23 <SergeyLukjanov> bob_nettleton, ErikB, sorry, I'm wrong
18:18:35 <SergeyLukjanov> in this case - upgrading Ambari - it should be ok
18:18:48 <SergeyLukjanov> we're not storing any ambari version in the plugin
18:19:04 <tosky> bob_nettleton: about this, isn't a potential issue to have an ambari version intended for hdp 2.1 into the official 2.0.6 images? (I guess no, otherwise you wouldn't do it, but...)
18:20:14 <bob_nettleton> tosky, Ambari supports deploying multiple versions of HDP.  The version we're moving to (1.6.0) supports the previous versions of HDP, and will eventually allow us to support HDP 2.1 as well.
18:20:16 <SergeyLukjanov> bob_nettleton, ErikB, the only thing is that probably we should have a doc that will describe that after migrating sahara from Icehouse to Juno you'll need to manually upgrade Ambari on HDP clusters?
18:20:16 <ErikB> @tosky -Ambari supports multiple stacks. The upgrade is a pre-req for us moving to support HDP 2.1
18:20:32 <bob_nettleton> tosky, the images won't need to be modified for this change to work, although eventually I'll pre-install this version on Ambari on the HDP images.
18:21:35 <ErikB> SergeyLukjanov, we can make doc updates to specify the process to upgrade if there is anything in particular that needs to be done.
18:21:47 <tosky> bob_nettleton, ErikB: I see, thanks; just to be sure, does it mean that, as long as the sahara plugin requires 2.0.6, that version will be installed even with a newer ambari?
18:22:16 <SergeyLukjanov> ErikB, I mean at least note that it should be done w/o instructions (link to ambari docs will be enough)
18:22:56 <bob_nettleton> tosky, that's correct.  we'll still be able to install HDP 2.0.6 (and 1.3.2) after this change.  If we add support for newer stacks, we may eventually decide to deprecate older ones, but this change does not force that to occur.
18:23:05 <SergeyLukjanov> ErikB, bob_nettleton, anyway, this doc could be added in separate change, so, I think that we're unblocking this change and could merge it when review will be done finally
18:23:07 <ErikB> tosky, yes you will be able to install 2.0.6 and later on other versions 2.1.
18:23:37 <tosky> thanks!
18:24:16 <bob_nettleton> SergeyLukjanov, great. could we ask Sergey R. to remove his -1 then? I think it was added in error anyway.  Then, if we could get another core reviewer to look at the patch, that would be great as well.  :)
18:24:49 <SergeyLukjanov> sreshetnyak, ^^
18:24:57 <aignatov> sreshetnyak: here? please remove -1 :) ^^
18:25:00 <SergeyLukjanov> bob_nettleton, I'll take a look on this patch after the meeting
18:25:53 <sreshetnyak> removed :)
18:25:58 <bob_nettleton> SergeyLukjanov, thanks, I appreciate that.
18:26:01 <SergeyLukjanov> bob_nettleton, oops, it's already have +2 from me :)
18:26:12 <bob_nettleton> sreshetnyak, thanks.  :)
18:26:19 <SergeyLukjanov> bob_nettleton, so, aignatov / alazarev / mattf could help with it
18:26:32 <SergeyLukjanov> bob_nettleton, ErikB, will jspeidel return back?
18:27:36 <ErikB> SergeyLukjanov, he's on vacation at the moment, and we are still in planning stage.
18:27:45 <SergeyLukjanov> ErikBm oh, ok
18:28:44 <aignatov> SergeyLukjanov, bob_nettleton: I think alazarev should do this merge since he raised this topic ;)
18:28:51 <SergeyLukjanov> #agreed https://review.openstack.org/#/c/98260/ is ok, waiting for doc in separated patch about need to upgrade ambari while upgrading sahara
18:29:20 <SergeyLukjanov> anything more to discuss in this topic?
18:29:51 * mattf needs to respond to the ^^ related emails
18:30:00 <aignatov> someone should reply in ML :)
18:30:31 <aignatov> mattf thinks the same ;)
18:30:35 <alazarev> SergeyLukjanov: sure
18:31:07 <SergeyLukjanov> aignatov, I'll floow-up with meeting agreement to the thread
18:31:26 * mattf loans his brain&voice to aignatov
18:31:30 <SergeyLukjanov> any comments on this topic are very welcome (to the thread too)
18:31:58 <SergeyLukjanov> okay, let's move on
18:32:00 <SergeyLukjanov> #topic Pilot sahara-specs
18:32:16 <SergeyLukjanov> we have the first one merged
18:32:17 <SergeyLukjanov> #link https://github.com/openstack/sahara-specs/blob/master/specs/juno/edp-move-examples.rst
18:32:22 <SergeyLukjanov> thanks to tmckay
18:32:41 <SergeyLukjanov> and we have action items for at least three more specs
18:32:45 <aignatov> mattf: ok :) one brain is good, two brains are much more better ;)
18:33:00 <SergeyLukjanov> aignatov, and multibrain is awesome
18:33:01 <elmiko> SergeyLukjanov: once i have a spec for this security related bug, how do we want to handle it?
18:33:09 <SergeyLukjanov> especcially with shared cache
18:33:24 <SergeyLukjanov> elmiko, it's a good question
18:33:30 <mattf> pinky&brain is best
18:33:38 <SergeyLukjanov> mattf, exactly!
18:33:43 <elmiko> lol
18:34:33 <SergeyLukjanov> elmiko, it's not a surprising issue and all users know about it, so, probably when we'll have a solution for it - it could be openede
18:34:44 <SergeyLukjanov> elmiko, at least to collect feadback for this solution
18:34:58 <elmiko> SergeyLukjanov: ack
18:35:39 <mattf> it sounds like we have an improved position (use token, don't store password) readily available and an enhancement (narrow token) in the future
18:36:04 <SergeyLukjanov> so, the pilot is going slow but good
18:36:14 <elmiko> and another enhancement further down the road, re: barbican
18:36:38 <mattf> yeah, we discussed how we cannot have a dep on a non-integrated project
18:36:57 <elmiko> right, that would be like for k release
18:37:12 <SergeyLukjanov> elmiko, yeah
18:37:46 <SergeyLukjanov> but this timing is looks ok if will take in account time for implementing all the stuff )
18:39:02 <elmiko> i'll proceed on making the spec, then we can talk about posting it to the git repo
18:39:54 <mattf> +1
18:39:57 <SergeyLukjanov> elmiko, agreed
18:40:38 <SergeyLukjanov> okay, let's move on
18:40:58 <SergeyLukjanov> There're no proposed changes for Roadmap update / cleanup
18:41:07 <SergeyLukjanov> so, let's do
18:41:09 <SergeyLukjanov> #topic Open discussion
18:42:41 <aignatov> lets fill the void :)
18:42:51 <SergeyLukjanov> yeha
18:42:53 <SergeyLukjanov> yeah*
18:42:58 <SergeyLukjanov> I'm a bit sleepy today
18:43:13 <SergeyLukjanov> heh, "a bit" -> "a lot"
18:43:35 <mattf> is someone +A'ing the hdp review?
18:44:11 <SergeyLukjanov> mattf, feel free to do it, it was already for enough time on review
18:44:21 <mattf> will do
18:44:33 <bob_nettleton> mattf, thanks!
18:44:39 <SergeyLukjanov> and looks like alazarev isn't here atm :)
18:44:52 <alazarev> returned
18:45:06 <mattf> done
18:45:06 <alazarev> air moved to reboot, strange
18:45:10 <aignatov> what’s about RobLevas’s patch? I’m ok with his changes too
18:45:20 <SergeyLukjanov> aignatov, link?
18:45:26 <RobLevas> aignatov: thanks.
18:45:39 <aignatov> https://review.openstack.org/#/c/101050/
18:46:08 <aignatov> wohooo
18:46:26 <crobertsrh> same thing happened last week
18:46:50 <mattf> not again!
18:47:00 <tosky> on another freenode channel I've seen this global notice:-tomaw- [Global Notice] Hi all. I'm going to reroute a hub and then restart a client server for a few updates. It'll be noisy but should be brief!
18:47:55 <SergeyLukjanov> hehe freenode makes us "separated"
18:48:13 <alazarev> :)
18:48:38 <aignatov> I’m still alive :)
18:49:46 <aignatov> I think we can end this meeting
18:50:06 <mattf> +1
18:50:14 <elmiko> +1
18:50:39 <aignatov> -2 [Do not merge], hehehe
18:50:44 <aignatov> :)
18:50:49 <elmiko> lol
18:51:01 <crobertsrh> heh
18:51:25 <mattf> aignatov, so old skool of you, it's -2 workflow now!
18:51:31 <mattf> err workflow-2
18:51:43 <SergeyLukjanov> mattf, it's a WIP state
18:51:52 <SergeyLukjanov> mattf, we still have -2 (code review)
18:52:15 <mattf> it's not well differentiated in the auto generated emails, that's for sure
18:52:18 <SergeyLukjanov> okay, so, if there are no other topics to talk about - I'll end the meeting in 3 mins
18:52:29 <RobLevas> Before you kill the meeting... did you decide on moving my patch along?
18:52:35 <RobLevas> https://review.openstack.org/#/c/101050/
18:53:20 <aignatov> RobLevas: you’ve got my +2 :)
18:53:22 <mattf> i'll take a look
18:53:33 <aignatov> but I reviewed it earlier
18:53:34 <RobLevas> aignatov: thank!
18:53:38 <aignatov> looks good
18:53:41 <RobLevas> mattf: thanks
18:53:42 <SergeyLukjanov> RobLevas, I'll take a look on it too
18:53:51 <SergeyLukjanov> and I'm ok with the blueprint RobLevas
18:53:52 <aignatov> I think toher guys still need to look at it
18:53:52 <RobLevas> SergeyLukjanov: thanks.
18:54:11 <RobLevas> SergeyLukjanov: I can add more info if you want.
18:54:12 <aignatov> toher - other
18:54:38 <SergeyLukjanov> RobLevas, it'll be great if you're write a spec for it (review to sahara-specs)
18:54:39 <RobLevas> I'm glad no one dinged me on the size of the patch. :)
18:54:59 <SergeyLukjanov> RobLevas, it'll be a good example for service-adding spec
18:55:02 <mattf> RobLevas, just you wait
18:55:21 <aignatov> RobLevas: next time you should split such patches logically into smaller ones :)
18:55:25 <mattf> SergeyLukjanov, +1 let spec slide atm, catch it for the next time around
18:55:26 <RobLevas> SergeyLukjanov: I will look into that.  I need to figure out what that all means. :)
18:55:31 <RobLevas> mattf: :)
18:55:38 <mattf> aignatov, +1
18:55:50 <RobLevas> aignatov: I now know for the future
18:56:18 <mattf> RobLevas, this isn't apache hadoop ya know, we like small changes that haven't been in progress for months...
18:56:31 <SergeyLukjanov> mattf, yeah
18:56:44 <SergeyLukjanov> and we're not using Jira to publish diffs for review
18:56:50 <SergeyLukjanov> so, it's not so painful :)
18:57:13 <mattf> also the turnaround on patches should be fast enough to make smaller patches viable
18:57:15 <SergeyLukjanov> oh no, netsplit again?
18:57:31 <RobLevas> mattf: The change isn't as big as the number of lines of code make it out to be... in my opinion. :)
18:57:32 <aignatov> small patches like this https://review.openstack.org/#/c/102658/ will work for me ;)
18:58:06 <SergeyLukjanov> aignatov, +2'd ;)
18:58:08 <RobLevas> aignatov: :)
18:58:21 <SergeyLukjanov> okay, thank you folks!
18:58:25 <SergeyLukjanov> #endmeeting