17:59:45 <SergeyLukjanov> #startmeeting sahara
17:59:46 <openstack> Meeting started Thu Jul 31 17:59:45 2014 UTC and is due to finish in 60 minutes.  The chair is SergeyLukjanov. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:59:47 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:59:49 <openstack> The meeting name has been set to 'sahara'
18:00:58 <crobertsrh> Hello
18:01:01 <SergeyLukjanov> let's wait for 5 mins for other folks
18:01:51 <ylobankov> hi
18:02:01 <aignatov> hello
18:02:11 <tmckay> o/
18:02:11 <alazarev> o/
18:02:21 <reshetnyaks> hi
18:02:22 <tmckay> \o
18:02:29 <SergeyLukjanov> okay, let's start the meeting
18:02:39 <SergeyLukjanov> #link https://wiki.openstack.org/wiki/Meetings/SaharaAgenda
18:02:45 <SergeyLukjanov> #topic News / updates
18:02:48 <SergeyLukjanov> folks, please
18:03:09 <tosky> hi
18:03:23 <tmckay> been working on specs and blueprints, more spark stuff -- spark job type, spark tests
18:03:26 <crobertsrh> I'm making progress on Juno dashboard work, https://etherpad.openstack.org/p/sahara-juno-post-merge-changes is my etherpad for keeping track.
18:03:35 <tmckay> we also need to figure out what to do about job cancel/delete
18:03:45 <alazarev> I’m implemented security group management feature, everyone are welcome to review
18:04:05 <tosky> so, just a note to say that a) the new CentOS images works b) the fix from Hortonworks to their ambari repository works too; I generated CentOS vanilla2 and HDP2 images
18:04:13 <elmiko> i've been working on the swift/auth blueprint, i have a few reviews up currently that could use more eyes(thanks to aignatov and alazarev). i'm working on the periodic task and trusts for job executions now.
18:04:18 <tosky> (funnily enough, with master dib, it fails on the stable 0.1.17)
18:04:28 <reshetnyaks> I'm working on CDH plugin and various bug fixing
18:04:30 <SergeyLukjanov> tosky, great
18:04:35 <SergeyLukjanov> (re centos images)
18:04:37 <alazarev> also fixed a bug in horizon that caused broken error messages in our dashboard
18:04:43 <crobertsrh> Side note:  The dashboard in horizon looks ugly right now due to their merging of the bootstrap library upgrade.  I have one patch that takes care of the serious breakages, but there are still some visual oddities all over the dashboard, not just our stuff.
18:04:50 <elmiko> tosky: we need to update the default tag once dib gets that extlinux patch tagged
18:04:56 <tosky> elmiko: definitely
18:04:57 <NikitaKonovalov> I've added first benchmark to rally testing sahara node group templates
18:05:15 <SergeyLukjanov> crobertsrh, NikitaKonovalov started looking on it too
18:05:18 <tosky> if anyone knows how to push a patch in a faster way into dib, please help :)
18:05:28 <SergeyLukjanov> elmiko, ++
18:05:32 <aignatov> no special from me this week
18:05:53 <SergeyLukjanov> tosky, send me a link to the patch and I ask 3o folks to review
18:06:02 <elmiko> SergeyLukjanov: i've just been waiting for the dib guys to get that patch in and make the next tag before i make a new PR
18:06:19 <reshetnyaks> please, review CDH plugin
18:06:35 <reshetnyaks> #link https://review.openstack.org/#/c/107390/
18:06:43 <SergeyLukjanov> reshetnyaks, I've +2'd it today
18:06:45 <alazarev> crobertsrh: is it Ok that order of our menu is random? or is it designed like it is now?
18:06:50 <tosky> SergeyLukjanov: https://review.openstack.org/#/c/104136/
18:07:00 <SergeyLukjanov> there was a -1 from tmckay for cdh, tmckay could you please take a look on it again?
18:07:04 <SergeyLukjanov> tosky, ack
18:07:09 <tosky> SergeyLukjanov: thanks
18:07:09 <tmckay> yes
18:07:13 <crobertsrh> alazarev:  Not random, but not of any particular design either.
18:07:50 <crobertsrh> I'd probably +1 just about any change to the order of panels :)
18:07:51 <aignatov> let’s fleshmob on tosky patch ;)
18:07:51 <reshetnyaks> SergeyLukjanov, need more reviews
18:08:16 <tmckay> lol, flashmob
18:08:23 <alazarev> crobertsrh: are we going to fix that? I was trying to find ‘Clusters’ that is somewhere in the bottom now
18:08:26 <tmckay> fleshmob might be something else
18:08:54 <SergeyLukjanov> alazarev, share screenshot ;)
18:09:04 <crobertsrh> I can add a bug to re-order things.  What is your proposed order?  Maybe the same order we had in sahara-dashboard?
18:09:21 <crobertsrh> I think I had put thought into the sahara-dashboard ordering, but I'm open to ideas
18:09:23 <SergeyLukjanov> crobertsrh, I think the same order as in s-d is preferd
18:09:24 <aignatov> tmckay: ops, please explain me what it could be private if it’s something awful…
18:09:31 <crobertsrh> ok
18:09:54 <tmckay> aignatov, lol, I have no idea :)  zombies? just a joke
18:10:14 <SergeyLukjanov> any other news/updates?
18:10:23 <elmiko> tmckay: we know what you were thinking... ;)
18:10:37 <SergeyLukjanov> #topic sahara-dashboard @ horizon status (croberts)
18:10:45 <SergeyLukjanov> #link https://etherpad.openstack.org/p/sahara-juno-post-merge-changes
18:11:01 <SergeyLukjanov> I'd like to keep this topic to be in track with our dashboard part ;)
18:11:09 <crobertsrh> Good idea
18:11:26 <crobertsrh> Hopefully, the cosmetic issues will be fixed soon
18:11:27 <SergeyLukjanov> I have an issue with internet connection
18:11:36 <SergeyLukjanov> how many patches are not landed yet?
18:12:01 <crobertsrh> I have 4 patches that are outstanding, plus I think alazarev has one or two
18:12:10 <NikitaKonovalov> there is a new bug report for dashboard https://bugs.launchpad.net/sahara/+bug/1349807
18:12:15 <uvirtbot> Launchpad bug 1349807 in sahara "[UI] Failed to copy cluster template" [Undecided,New]
18:12:26 <NikitaKonovalov> probably need a fix to both horizon and s-d
18:12:54 <NikitaKonovalov> looks like it affects stable/icehouce also
18:13:01 <SergeyLukjanov> oh
18:13:56 <SergeyLukjanov> #info sahara@horizon status is green
18:14:32 <SergeyLukjanov> #topic Juno 3 (Sep 4)
18:14:43 <SergeyLukjanov> #link https://wiki.openstack.org/wiki/Juno_Release_Schedule
18:14:51 <alazarev> are we going to keep dashborad for juno?
18:15:01 <SergeyLukjanov> #info j3 is FF and SF
18:15:03 <crobertsrh> I should note that the horizon people appear to be bringing their feature freeze date up a bit.
18:15:13 <SergeyLukjanov> alazarev, nope, I don't see any issues
18:15:34 <crobertsrh> They are proposing a freeze in just a couple of weeks.  The rationale is so that they have time to review everything.
18:15:44 <SergeyLukjanov> crobertsrh, I'll talk with David about their police
18:15:53 <crobertsrh> Great
18:16:00 <alazarev> police :)
18:16:03 <SergeyLukjanov> but AFAIK they are accepting changes for a bit more time than other projects
18:16:15 <SergeyLukjanov> police -> policy
18:16:54 <SergeyLukjanov> #action SergeyLukjanov to talk with Horizon folks about features merge deadline
18:16:59 <aignatov> dashboard police :) nice :)
18:17:08 <SergeyLukjanov> #topic Open discussion
18:17:25 <SergeyLukjanov> #info sahara-specs is growing very well and that's cool
18:17:30 <tmckay> job deletion semantics, and exposing job cancelation in the client and UI
18:17:36 <tmckay> my pet issue ^^
18:17:39 <elmiko> do we have anything like a singleton pattern in Sahara?
18:17:51 <SergeyLukjanov> tmckay, yeah, we should discuss it
18:17:55 <tmckay> elmiko, I think there is one somewhere
18:18:48 <elmiko> i'm still trying to figure out the best way to maintain this global information without commiting to the db
18:18:49 <tmckay> Summary:  I think delete should be "cancel if runnig then delete record from the db" and cancel is just cancel.  Cancel is implemented, and available from REST, but not exposted otherwise
18:18:55 <SergeyLukjanov> tmckay, could you please explain the issue to make everyone understand
18:18:57 <SergeyLukjanov> tmckay, :)
18:18:57 <tmckay> delete is just remove record from the db
18:19:06 <tmckay> job executions, this is ^^
18:19:29 <alazarev> elmiko: new db engine had facade as singletone
18:19:31 <tmckay> so, currently, you cannot stop an EDP job
18:19:41 <tmckay> just erase the record for it -- it keeps running
18:19:55 <tmckay> "cancel" is good because you can stop and relaunch
18:20:18 <aignatov> tmckay: agree, +2 :)
18:20:25 <tmckay> for ephemeral cluster, it doesn't matter so much
18:20:26 <elmiko> tmckay: agree about splitting to cancel and delete
18:20:41 <elmiko> alazarev: thanks, does that store to the db?
18:21:03 <tmckay> so, to expose it, we need new client method (doesn't break anything, just addition), and a new UI button
18:21:15 <SergeyLukjanov> tmckay, +2
18:21:28 <aignatov> I thought we already have the same functionality and behaviour you are talking about
18:21:36 <tmckay> and deletion semantics change to "attempt cancel, then delete".  I think erasing a running job exec without cancel is silly -- it's running, you just can't see it
18:21:46 <tmckay> aignatov, nope
18:21:53 <tmckay> sad story
18:21:57 <crobertsrh> +1, seems like the right thing to do
18:21:58 <tmckay> oversight, I think
18:22:08 <tmckay> Okay, I'll finish the spec
18:22:11 <alazarev> elmiko: no, it is for facade, just to create it once
18:22:36 <elmiko> alazarev: ok, i'll have to investigate. that may work for what i want.
18:22:46 <SergeyLukjanov> tmckay, agreed
18:23:18 <tmckay> #action tmckay finishes spec and blueprint for job execution cancel/delete semantics
18:24:13 <tmckay> oh, SergeyLukjanov
18:24:16 <tmckay> Paris
18:24:19 <tmckay> live demo?
18:24:19 <SergeyLukjanov> do we have anything more to discuss?
18:24:23 <SergeyLukjanov> tmckay, sure ;)
18:24:29 <tmckay> I am afraid :)
18:24:38 <tmckay> but the talk proposal says we'll do it
18:24:44 <SergeyLukjanov> tmckay, I think we could ask dmitryme to prepare it talk will be accepted
18:25:17 <SergeyLukjanov> tmckay, we could remove it from talk proposal if we'll be too scared
18:25:18 <tmckay> okay. I am worried about trying again with Linux laptop, and connecting back to RH cluster
18:25:28 <tmckay> might not be fast enough
18:25:38 <tmckay> and it might work better from a Mac :)
18:25:47 <tmckay> maybe a Mac to a Mirantis cluster?
18:25:54 <SergeyLukjanov> tmckay, it defenetly works better from Mac ;)
18:26:03 * tmckay hangs head in shame
18:26:05 <SergeyLukjanov> tmckay, dmitryme is Mac user too
18:26:18 <crobertsrh> pretty lame tmckay
18:26:24 <tmckay> do you guys have infrastructure for a fast cluster near Paris?
18:26:29 <SergeyLukjanov> heh, definitely*
18:26:38 <tosky> tmckay: ehm
18:26:50 <tmckay> tosky, yes?
18:26:50 <SergeyLukjanov> tmckay, I think so, we have some hardware in Europe
18:26:58 <tosky> tmckay: I won't comment further :D
18:27:01 <aignatov> our CI running in Czech
18:27:09 <alazarev> we has lab in czech, it is pretty near
18:27:13 <SergeyLukjanov> tmckay, and we have a very performant MBP 15" that could run clusters and exec jobs
18:27:16 <tmckay> tosky, I'll talk to you offline
18:27:40 <tmckay> okay, sounds good then.  We should think about apps.  Maybe a cool spark app
18:28:19 <aignatov> bigpet store in Spark?;)
18:28:48 <tmckay> I should note for the record, the project broke in Atanta, it was not Linux's fault
18:28:49 <elmiko> aignatov: it hasn't been rewritten for spark yet
18:28:52 <tmckay> Up with Linux
18:29:09 <tmckay> I can't help it if the projector was a Mac-lover
18:29:23 <tmckay> projector broke, that is
18:29:31 <elmiko> it's an ops issue
18:30:07 <SergeyLukjanov> tmckay, yeah, it was a projector failure
18:31:43 <SergeyLukjanov> tmckay, btw dmitryme is on PTO now, so, that's why I'm "sure" he'll prepare demo :)
18:31:52 <elmiko> LOL
18:32:30 <tmckay> lol, great
18:32:58 <SergeyLukjanov> anything else or 28 free mins?
18:33:13 <crobertsrh> Nothing from me
18:33:33 <elmiko> please take a look at https://review.openstack.org/#/c/110134/ and https://review.openstack.org/#/c/110841/
18:33:45 <elmiko> thanks to those who have :)
18:34:12 <SergeyLukjanov> okay, thanks folks!
18:34:15 <SergeyLukjanov> #endmeeting