18:07:28 <SergeyLukjanov> #startmeeting savanna
18:07:29 <openstack> Meeting started Thu Jan 23 18:07:28 2014 UTC and is due to finish in 60 minutes.  The chair is SergeyLukjanov. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:07:30 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
18:07:32 <openstack> The meeting name has been set to 'savanna'
18:08:01 <SergeyLukjanov> #topic Agenda
18:08:16 <SergeyLukjanov> #link https://wiki.openstack.org/wiki/Meetings/SavannaAgenda#Agenda_for_January.2C_23
18:08:37 <SergeyLukjanov> action items are partially resolved
18:08:47 <SergeyLukjanov> will make an update next time
18:08:52 <SergeyLukjanov> #topic Icehouse-2 dev milestone
18:09:26 <SergeyLukjanov> #link http://lists.openstack.org/pipermail/openstack-dev/2014-January/025250.html
18:09:36 <SergeyLukjanov> you can find links and details in this email
18:09:59 <SergeyLukjanov> we've successfully delivered icehouse-2 with mostly all planned features and bugs fixed
18:10:12 <SergeyLukjanov> so, my congratulations!
18:10:17 <mattf> congrats all!
18:10:39 <ErikB> Great job everyone!
18:10:55 <SergeyLukjanov> it's our second dev milestone
18:11:01 <mattf> it's terrific to see all the progress on heat integration
18:11:02 <aignatov> hurrrah!
18:11:21 <alazarev> congrats!
18:11:25 <SergeyLukjanov> it was handled by ttx
18:11:32 <SergeyLukjanov> for the main savanna project
18:11:56 <SergeyLukjanov> ok, let's move on
18:12:06 <SergeyLukjanov> #topic News / updates
18:12:10 <SergeyLukjanov> folks, please
18:12:14 <SergeyLukjanov> share news
18:12:29 <SergeyLukjanov> preferably good news, because I have some bad ones
18:12:43 <crobertsrh> UI part of java actions and HDFS are up for review, seem positive so far.
18:12:47 <mattf> just a few -create cli commands short of having full coverage of the v1.1 api
18:12:51 <ErikB> On the HDP side, we are moving forward with EDP change, Blueprints, DiskBuilder and docs
18:12:57 <tmckay> I'm looking at more Oozie actions now, streaming mapreduce first
18:13:41 <dmitryme> I'am replacing SSH configuring with the guest agent
18:13:55 <aignatov> continue working on Heat integration polishing, fixing some bugs, also yesterday I've realised that savanna heat engine does;t working with heat in master
18:14:23 <SergeyLukjanov> ErikB, that's awesome to hear about dib elements for HDP
18:14:40 <aignatov> because of https://bugs.launchpad.net/heat/+bug/1271597
18:14:47 <alazarev> I'm working on blogpost about data locality support, hopefully will be published soon
18:14:49 <ErikB> SergeyLukjanov, will be a good change and Bob should me merging shortly.
18:15:01 <aignatov> but seems the bug will be fixed in the heat side soon
18:15:22 <bob_nettleton> I hope to submit a patch for the HDP DIB elements sometime over the next few days.
18:15:37 <aignatov> bob_nettleton: awesome!!
18:15:42 <SergeyLukjanov> bob_nettleton, awesome, ping me or mattf if any questions
18:15:51 <bob_nettleton> great. thanks!
18:16:01 * mattf gets his meat grinder ready
18:16:08 <ErikB> SergeyLukjanov, what is the bad news…
18:16:25 <SergeyLukjanov> the bad news are about Savanna project naming
18:16:40 <mattf> forcing us to add a trailing h?
18:16:47 <ErikB> ;-)
18:16:53 <dmitryme> mattf, lol
18:16:57 <SergeyLukjanov> I've contacted OpenStack Foundation marketing team to ensure that we're using consumable names
18:17:04 * mattf hands out Gs Ns and Us
18:17:15 * SergeyLukjanov looking for the bad link
18:17:38 <SergeyLukjanov> aaaaand
18:17:41 <SergeyLukjanov> #link http://www.thetus.com/savanna
18:17:50 <mattf> other than savannah, we're almost past out painful name collision with "havana"
18:18:16 <SergeyLukjanov> mattf, yep, but it's not so bad
18:18:19 <aignatov> :-( so we should rename our wonderful project Savanna, right?
18:18:26 <SergeyLukjanov> aignatov, yep
18:18:35 * mattf groans
18:18:52 <SergeyLukjanov> and we should at least find the new name in next 2-4 weeks
18:19:03 <SergeyLukjanov> to be able to rename the world before the graduation review
18:19:14 <mattf> SergeyLukjanov, have a pointer to the trademark registration?
18:19:43 <aignatov> proposing new name - Saratov :)
18:19:46 <SergeyLukjanov> mattf, their project was released first time Feb 2010
18:19:58 <SergeyLukjanov> and so it's at least bad to use the same name
18:20:08 <SergeyLukjanov> because it's a project in the same area - cloud, hadoop
18:20:11 <DinaBelova> SergeyLukjanov you may combine the fact you need new name with the fact you are not only Hadoop now - you are "Data processing" OS program
18:20:42 <DinaBelova> SergeyLukjanov, and elephants topic (Savanna) seems to be not right now, really :)
18:20:50 <SergeyLukjanov> that's the only pros I see...
18:20:55 <DinaBelova> so maybe renaming is a good idea anyway
18:21:06 <SergeyLukjanov> nope, I really love the Savanna naming!
18:21:12 <DinaBelova> :D
18:21:20 <aignatov> it could be an elephant topic still, why not?
18:21:23 <alazarev> we can use 'Savannah' :D
18:21:30 <SergeyLukjanov> I dislike h
18:21:30 <mattf> SergeyLukjanov, do they have a trademark on the name Savanna, is the os foundation going to acquire a trademark on our name?
18:21:43 <SergeyLukjanov> mattf, I'll clarify that
18:21:53 <tmckay> How about "Grassland" ? :)
18:22:19 <SergeyLukjanov> :)
18:22:39 <tmckay> StackVanna?
18:22:45 <aignatov> lol!
18:22:56 <SergeyLukjanov> I'll start a ML thread when ensure all details
18:23:05 <SergeyLukjanov> and we'll try to find the better name
18:23:08 <SergeyLukjanov> than npthing ;)
18:23:28 <SergeyLukjanov> I hope that will find the great naming that'll be ok for all ofus
18:23:39 <alazarev> we should at least 1. check is Savanna is really registered 2. contact thetus and try to buy name
18:23:41 <mattf> is there an existing trademark on Savanna? is the openstack foundation going to trademark our eventual name?
18:24:14 <SergeyLukjanov> mattf, I'll ask foundation about that
18:24:19 <SergeyLukjanov> don't worry :)
18:24:36 <mattf> this is going to be especially rough on ErikB, who still uses quantum
18:24:57 <mattf> ErikB, we're here to help
18:25:52 <DinaBelova> Even Google don't know about these Thetus Savanna :) Try "cloud hadoop savanna" - you'll see only your Savanna
18:26:32 <DinaBelova> don't -> doesn't
18:26:34 <mattf> this stuff is best left to the lawyers, but we definitely should jump ship on the name unless we have a solid reason
18:26:43 <SergeyLukjanov> DinaBelova, that's right, but it looks like it'll not help us :(
18:27:26 <SergeyLukjanov> heh, so, to summarize
18:27:47 <SergeyLukjanov> I'll talk again with foundation folks to clarify stuff
18:28:03 <SergeyLukjanov> and than start a mailing thread (if needed) to find the name
18:28:12 <SergeyLukjanov> let's move on
18:28:16 <SergeyLukjanov> #topic Meeting re-org
18:28:31 <SergeyLukjanov> I was thinking about how our meeting working now and have some ideas
18:28:43 <SergeyLukjanov> I'd like to share them and receive some feedback
18:29:07 <SergeyLukjanov> #1 add topics with topic chairs how will be responsible for them
18:29:25 <SergeyLukjanov> like Heat integration update or CI status
18:29:41 <aignatov> how->who?
18:29:49 <SergeyLukjanov> yep
18:30:01 <SergeyLukjanov> it could contain about 1-2 sentences with short clearly defined status news
18:30:05 <SergeyLukjanov> status/news
18:30:28 <SergeyLukjanov> idea #2:
18:30:39 <SergeyLukjanov> heh, I completely forgot the second idea
18:30:49 <SergeyLukjanov> any thoughts on the first one?
18:30:54 <mattf> i guess we can settle #2 then
18:31:09 <SergeyLukjanov> :)
18:31:23 <mattf> #1 could just be a rundown of active bps, each has an assignee
18:31:38 <aignatov> what is the difference between idea1 and what is happening right now when we just say news/updates?
18:31:57 <mattf> i think we get the information into the minutes anyway. if there's value in structuring it more, i'm ok witht hat
18:32:41 <aignatov> Chad works on the UI, he updates us about it, Trevor owirks on EDP, he updates us, I work on Heat, I update you
18:33:01 <aignatov> owirks -> works :)
18:33:14 <SergeyLukjanov> aignatov, that's correct
18:33:24 <SergeyLukjanov> the idea was to ensure that all parts are covered
18:33:26 <alazarev> I also don't see difference with what we have right now
18:33:35 <SergeyLukjanov> for example I'd like to see CI status updates
18:34:06 <mattf> afaict #1 is some added structure, seems low overhead, but the added value hasn't been articulated yet
18:34:06 <SergeyLukjanov> probably structuring is overcomplexity
18:34:16 <alazarev> SergeyLukjanov: if you like CI updates - you can always ask about it
18:34:47 <SergeyLukjanov> alazarev, I'd like to avoid pings, just short updates
18:35:03 <SergeyLukjanov> because we already have enough stuff that could be shared weekly
18:35:57 <SergeyLukjanov> so, currently it looks for me that we could distribute interesting topics between folks and everyone will now what should he/she cover
18:36:08 <SergeyLukjanov> and will be able to prepare
18:36:10 <SergeyLukjanov> for ex.
18:36:33 <SergeyLukjanov> if you really  dislike such idea
18:36:41 <SergeyLukjanov> write about it :)
18:37:37 <SergeyLukjanov> any thoughts?
18:38:09 <mattf> i'm ambivalent
18:38:38 <ErikB> same here
18:38:39 <alazarev> sounds reasonable, the only thing needed is a list of topics with assignments, don't think that we need some kind of ball passing
18:38:50 <jmaron> I'm ambi-normal
18:38:54 <crobertsrh> Seems like it's worth a shot.  Maybe it will help someone with something.  If not, we can always change back.
18:39:28 <aignatov> well, I still do not see the much difference, but we can try, maybe I didn't get your idea
18:39:30 <SergeyLukjanov> alazarev, sure, w/o ball passing, just updates
18:39:56 <SergeyLukjanov> aignatov, I'd like to ensure that we have an aggregated weekly updates
18:40:16 <SergeyLukjanov> personally I think that it's awesome to have a weekly updates in meeting logs
18:40:43 <SergeyLukjanov> any thoughts on topics that should be covered?
18:40:46 <alazarev> as I understand the difference, there will be update on CI even if there are no much changes in CI
18:40:59 <alazarev> SergeyLukjanov: is my understanding right?
18:41:02 <aignatov> alazarev: exactly
18:41:08 <ErikB> it would be good to get update on 'active' blueprints
18:41:21 <SergeyLukjanov> alazarev, I'd like to see some stats for CI for example
18:41:28 <SergeyLukjanov> from CI guys
18:41:34 <mattf> ErikB, nice idea
18:41:42 <SergeyLukjanov> ErikB, agreed
18:41:44 <aignatov> ErikB: good point
18:41:54 <jmaron> dependency updates.  I stumbled (painfully) across the new need for mysql and postgres installs on my dev obx
18:41:56 <jmaron> box
18:42:13 <alazarev> so, we just need to document list of topics somewhere
18:42:26 <alazarev> not much changes to existing structure
18:42:31 <mattf> i painfully stumbled on use_namespaces requiring savanna-api be run as root, oof
18:42:50 <jmaron> that wasn't that painful… ;)
18:43:19 <SergeyLukjanov> I'll try to list topics to the next meeting
18:43:21 <mattf> maybe not for you!
18:43:28 <SergeyLukjanov> let's move on to the open discussion
18:43:32 <jmaron> (btw, any luck corralling neutron experts)
18:43:37 <SergeyLukjanov> #topic General discussion
18:43:39 <mattf> some
18:43:56 <mattf> general topic from last week - everyone wants the cli to support IDs and NAMEs
18:44:01 <jmaron> I would hope there's a way to configure the nature of the namespace creation
18:44:11 <mattf> everyone also wants the savanna api (rest) to support them so the client can be thin
18:44:19 <SergeyLukjanov> mattf, btw, how is your cli work? do you have estimates to name it beta cli?
18:44:29 <mattf> i did some digging and it looks like we may be unique in having that functionality in our rest api
18:44:48 <jmaron> trail blazers!
18:44:52 <SergeyLukjanov> mattf, all projects support it on the client side
18:44:53 <mattf> sooooo... how strongly do people want the client to be thin?
18:45:05 <mattf> SergeyLukjanov, yeah, where were you last week to tell us that!?
18:45:09 <SergeyLukjanov> because mostly all other project have no constraint on resource name
18:45:20 * mattf has since learned this
18:45:23 <SergeyLukjanov> I was on the PTL's webinar :)
18:45:40 <mattf> excuses excuses
18:46:14 <mattf> so, unless someone who hasn't been hitting the hwx bar pressures me really hard, the name support is going to be client side - at least for the first pass i write
18:46:30 <SergeyLukjanov> at least for the v1.1 api
18:46:56 <mattf> yeah, i dunno what magical support will be unlocked with the new v2 impl that may use an entirely different framework
18:47:27 <SergeyLukjanov> mattf, I'd like to not harry with v2 api and include all nice features to it
18:47:38 <SergeyLukjanov> that all of us dreaming about ;)
18:47:57 <mattf> that's nice, we can have a dv2 (dream v2) too
18:48:04 <SergeyLukjanov> like /v2/beer/one-more-please
18:48:13 <mattf> in the meantime, i'll be busy on v2
18:48:36 <SergeyLukjanov> mattf, are you planning to complete your 'run_tests.sh'?
18:48:44 <mattf> sounds like no objections to name search on the client side, thank!
18:48:59 <SergeyLukjanov> mattf, not from my side
18:49:05 <mattf> SergeyLukjanov, i'll take a look again this week, but it was still busted for savanna
18:49:24 <alazarev> mattf: name search on client side looks fine
18:49:33 <mattf> as for v1 and v1.1 - i'm a little disturbed that we store swift creds and pass them around
18:49:50 <mattf> i'm curious about how we happen to call them credentials in one call and extra in another though
18:49:54 <SergeyLukjanov> we need to add the corresponding properties to the filter
18:50:02 <SergeyLukjanov> to not return back them from the rest api
18:50:19 <aignatov> tmckay: I had a quick look at the savanna channel and observed that you plan to start working on the streaming api for edp
18:50:40 <mattf> data-sources uses credentials, job-binaries uses extra
18:51:19 <mattf> if no one speaks up, i'll probably propose them both as "credentials" for v2
18:51:25 <tmckay> aignatov, yes
18:51:28 <ErikB> +1
18:51:38 * mattf might even if folks speak up
18:51:41 <tmckay> I have a question about that, but I have to run to the doctor
18:52:05 <aignatov> tmckay: ok, we can talk later about it
18:52:07 <tmckay> aignatov, essentially, should streaming map-reduce be a new job type, or extra options for current map-reduce job type?
18:52:28 <tmckay> it effects crobertsrh, we should consider ease of use in the UI
18:52:29 <mattf> another oddity i found - clusters have a status property that provides you info about what you'd expect, cluster status
18:52:45 <mattf> job-executions have an info property that has a status property that tells you about the job-execution
18:52:48 <aignatov> tmckay: as I remember streaming is the subset of common MapReduce, right?
18:52:55 <alazarev> mattf: +1 on minimizing extra
18:52:56 <tmckay> yes
18:53:00 <mattf> anyone know why cluster.status and job-executino.info.status or know why they should be the same?
18:53:14 <mattf> alazarev, thx
18:53:21 <mattf> ErikB, thx
18:53:39 <mattf> s/why they should/why they should not/
18:53:41 <aignatov> tmckay: maybe in the UI side it should be as a separate job type but in the core EDP...
18:53:53 <SergeyLukjanov> mattf, I'd like to have j_e.status
18:53:54 <aignatov> how simple to implement :)
18:54:08 <tmckay> aignatov, could be, I had that thought to.  talk to you soon
18:54:09 <SergeyLukjanov> we have 6 mins
18:54:18 * mattf can be done
18:54:43 <aignatov> tmckay: not today, ok? tomorrow, ping me
18:54:53 <tmckay> ack
18:55:06 <alazarev> mattf: does job-executino.info contain something other than status?
18:55:22 <mattf> alazarev, afaik, no
18:55:45 <jmaron> 'info': {'status': 'Pending'},
18:56:04 <alazarev> mattf: may be something was supposed...
18:57:22 <mattf> ok, keep an eye out. i'll probably propose it gets collapsed for v2
18:58:33 <SergeyLukjanov> mattf, ok
18:59:02 <SergeyLukjanov> looks like we're out of time
18:59:06 <SergeyLukjanov> thank you all!
18:59:11 <SergeyLukjanov> have a good night/day
18:59:16 <SergeyLukjanov> #endmeeting