18:00:19 <SlickNik> #startmeeting trove
18:00:20 <openstack> Meeting started Wed Aug 13 18:00:19 2014 UTC and is due to finish in 60 minutes.  The chair is SlickNik. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:00:21 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
18:00:24 <openstack> The meeting name has been set to 'trove'
18:00:27 <denis_makogon> o/
18:00:34 <SlickNik> Agenda at:
18:00:36 <SlickNik> #link https://wiki.openstack.org/wiki/Meetings/TroveMeeting
18:00:44 <amcrn> o/
18:00:54 <kevinconway> o/
18:00:57 <pdmars> o/
18:01:01 <tvoran> o/
18:01:06 <grapex> o/
18:01:12 <amrith> ./
18:01:30 <SlickNik> #topic Juno Midcycle meetup proposed agenda
18:01:38 <SlickNik> #link https://wiki.openstack.org/wiki/Trove/JunoMidCycleMeetup
18:01:50 <iccha> o/
18:02:31 <vgnbkr> o/
18:03:02 <SlickNik> amrith: Did you have something for us?
18:03:06 <amrith> yes
18:03:10 <glucas> o/
18:03:11 <amrith> was waiting fr the hand-off ;)
18:03:17 <amrith> so a couple of things
18:03:26 <denis_makogon> it would never end =)
18:03:28 <dougshelley66> o/
18:03:29 <peterstac> o/
18:03:32 <schang> o/
18:03:41 <amrith> if you are coming in on Tuesday and are going to be at Trove day, there's a cocktail party in the evening
18:03:44 <amrith> please join us there
18:03:58 <amrith> On Wednesday evening we're going to have a happy hour.
18:04:07 <amrith> on Thursday there's a baseball game ...
18:04:16 <amrith> On Friday (unfortunately) we all go our merry ways ;)
18:04:26 <robertmy_> o/
18:04:28 <amrith> I will be emailing all attendees with location information
18:04:31 <amrith> for the meetings
18:04:38 <grapex> baseball game?
18:04:39 <amrith> I have not heard from anyone in response to my email
18:04:44 <grapex> and I just threw out my shoulder... :(
18:04:48 <amrith> asking about whether they'll be there or not
18:05:04 <amrith> so if you are not going to be attending (but had registered) please let me know
18:05:10 <amrith> so we don't waste food and stuff
18:05:16 <dougshelley66> grapex, i believe the baseball game is sponsored by Rax
18:05:22 <denis_makogon> amrith, what about hangout attendance?
18:05:23 <amrith> other than that, about the agenda
18:05:41 <nehav> that is correct rackspace is sponsoring the game
18:05:51 <nehav> and i have a count of 30 people for the game.
18:05:53 <amrith> it would be good if we could focus some time (maybe the 4 hours we have on bp and docs could give us some time) to talk about backlog
18:06:07 <amrith> thanks to rax for the game ...
18:06:12 <amrith> thanks to HP for the happy hour ...
18:06:20 <amrith> what else. I see some questions.
18:06:26 <denis_makogon> unfortunalety, i'm not able to come over, that's why i'd request the one of possible ways to participate through hangout/webex or skype
18:06:40 <amrith> will try to do hangout and things
18:06:47 <amrith> but the experience from austin was mixed
18:06:57 <denis_makogon> yeah
18:07:03 <denis_makogon> bad mic =(
18:07:06 <amrith> I'm not certain that we are in a location with high speed high bandwidth internet
18:07:13 <amrith> to support streaming video
18:07:30 <grapex> Plus a court order prohibits me from appearing on any videos on the internet
18:07:31 <denis_makogon> at least could someone record all days ?
18:07:40 <amrith> we had not planned on it
18:07:50 <amrith> that's a costly undertaking to get a professional quality video
18:08:01 <amrith> so if someone would like to volunteer to do it, that would be super.
18:08:13 <denis_makogon> it would be great to have video records in the end, to stay on the same way with all of you folks
18:08:14 <ruebenramirez> o/
18:08:29 <denis_makogon> another question is related to agenda
18:08:41 <glucas> grapex: we'll blur your face.
18:08:49 <cweid> o/
18:09:03 <denis_makogon> i don't see any topic related to monitoring thing
18:09:04 <amrith> SlickNik, ... I'm done.
18:09:52 <SlickNik> denis_makogon: It's under OpenStack integration
18:10:03 <SlickNik> (integration with ceilometer)
18:10:08 <SlickNik> thanks amrith
18:10:12 <denis_makogon> SlickNik, i see
18:10:15 <denis_makogon> thanks
18:10:16 <SlickNik> Any other questions?
18:10:20 <denis_makogon> nope
18:10:40 <iccha> SlickNik: will we have owners and etherpads for sessions like in oepnstack summit?
18:10:55 <iccha> and will they be up before the midcycle so we have an idea about what is going to be disucssed
18:11:01 <sgotliv> o/
18:11:05 <denis_makogon> i guess etherpads are reserve by the default
18:11:30 <SlickNik> iccha: Yes, I'll link the etherpads by today, so folks can go over and start filling them out over the next few days.
18:11:56 <SlickNik> iccha: Thanks for the reminder!
18:12:19 <SlickNik> Okay, let's move on to the next topic in the agenda.
18:12:36 <SlickNik> #topic Heads up: Clustering implementation has landed in Gerrit
18:12:45 <SlickNik> amcrn / mat-lowery: all yours
18:12:55 <amcrn> nothing much to say here, just an fyi that we've landed the first iteration of clustering in the public gerrit
18:13:07 <amcrn> working on porting an internal wiki with more details to a public one
18:13:29 <amcrn> but beyond that, let us know what you think, and ping myself or mat-lowery or ramashri if you have any questions
18:13:41 <denis_makogon> amcrn, cool, would be nice to see doc that describes implemented framework
18:13:58 <denis_makogon> amcrn, i've got question
18:14:24 <denis_makogon> amcrn, does clustering implementation for mongodb covers mongo replication ?
18:14:42 <amcrn> denis_makogon: are you referring to master/slave, or vanilla replica-sets, or ?
18:14:59 <denis_makogon> amcrn, correct, master-slave
18:15:00 <amcrn> either way, only sharding is support in this push.
18:15:05 <amcrn> supported*
18:15:17 <amrith> denis_makogon, I don't know what your question is about google calendars.
18:15:24 <denis_makogon> amcrn, i'd like to start working on mongo replication under replication v1
18:15:30 <amrith> denis_makogon, I don't use google calendar.
18:15:40 <denis_makogon> amrith, let's talk about that after the meeting
18:15:53 <amcrn> denis_makogon: let's chat about it offline, we've got some prior art for it
18:16:00 <amrith> denis_makogon, if it relates to mid-cycle, let's talk about it here.
18:16:04 <denis_makogon> amcrn, cool, thanks
18:16:12 <denis_makogon> amrith, ok
18:16:35 <denis_makogon> amrith, could you please create events in google calendar and share it for all ?
18:16:44 <amrith> denis_makogon, I don't use google calendar.
18:16:55 <kevinconway> amrith: sounds like you should check it out
18:17:12 <denis_makogon> amrith, it would be easier to use hangout and calendar events
18:17:12 <amrith> kevinconway, I will do that in my copious spare time. Maybe on 23rd ;)
18:17:37 <grapex> amrith: also, did you send out the save the dates? I never got mine. :(
18:17:43 <kevinconway> you can also have task checklists on your google calendar
18:17:46 <grapex> Now I don't know where Trove is registered!
18:17:48 <amrith> grapex, I save the date for you
18:18:03 <kevinconway> if you click on dates in your gmail it auto creates calendar events
18:18:10 <amrith> I saved the date for you, just for you.
18:18:11 <SlickNik> Okay, I think we're straying a bit off the topic here.
18:18:18 <amrith> SlickNik, sorry.
18:18:24 <denis_makogon> sorry
18:18:27 <kevinconway> you can store photos on google too!
18:19:01 <grapex> Let's all agree Google technology is interesting and move on. :p
18:19:08 <SlickNik> amcrn: Thanks for the heads up wrt the clustering patches!
18:19:14 <amcrn> SlickNik: word
18:19:25 <SlickNik> And for pushing them up.
18:19:28 <kevinconway> amrith: can you share the cluster patch with us on google plus?
18:19:37 <kevinconway> i'm done.
18:19:37 <amrith> +1
18:19:48 <glucas> kevinconway: LOL
18:19:53 <grapex> Sorry guys, kevinconway is actually a google stealth marketter
18:20:02 <SlickNik> #topic Open Discussion
18:20:10 <amrith> I have one ...
18:20:15 <denis_makogon> i've got one question
18:20:16 <nehav> i have one too
18:20:20 <denis_makogon> amrith, go ahead
18:20:33 <amrith> neha ... go ahead
18:20:33 <SlickNik> Okay, amrith go for it
18:20:37 <SlickNik> or neha
18:20:40 <amrith> chivalry isn't dead here ;)
18:21:41 <amrith> neha?
18:22:07 <amrith> I was only going to ask about this calendar thing. I'm not sure what this concept of calendar for hangouts is.
18:22:08 <nehav> cp16 proposed this bp in the public https://wiki.openstack.org/wiki/Trove/default-configuration-values-for-datastore-version-flavor
18:22:27 <nehav> for adding a new api call to create a config with default params
18:22:58 <nehav> we would like to add more information about the params and have extended the bp here https://wiki.openstack.org/wiki/Trove/verbose-default-configuration-values#Get_default_configuration_values_for_datastore_version_and_flavor_.28GET.29
18:24:07 <nehav> this would help that when we invoke this API call we can provide the min, max and other characteristics to the user which will be important. This would help users to be more aware of the param values and ensure they dont blow up the configs
18:24:14 <nehav> thoughts?
18:24:50 <amrith> neha, I haven't read it in detail but I will definitely provide feedback on Monday
18:24:58 <amrith> will it be on the agenda for the monday meeting?
18:25:27 <nehav> we can add it. i guess a lot of people may be travelling this mon though
18:25:42 <amrith> good point (some of us are in Boston already ;))
18:25:44 <SlickNik> So to understand correctly, this is in place of the earlier implementation? (or is it in addition to?)
18:26:13 <nehav> basically adding more info in the response of the api call
18:27:16 <cp16net> yeah adding the validation rules to each default config
18:27:20 <SlickNik> Got it. Seems reasonable that users would want to know the rule set for min value, max value, etc.
18:28:02 <SlickNik> I'd suggest giving folks some time to review the change, talking about it again on Monday (or at the mid-cycle)
18:28:03 <amrith> nehav, in effect you are asking for a way to get a listing of a configuration of a template directly, without having a running instance?
18:28:04 <nehav> yes it will be much easier to have the info in this api call, rather than calling config param details for each parameter
18:28:06 <cp16net> the payload changed from the original spec thats about all
18:28:14 <nehav> amrith correct
18:28:25 <cp16net> yup
18:28:31 <nehav> this is why this new api call was proposed
18:28:40 <amrith> the spec seems to focus on mysql. I assume some accommodation will be made for other datastores as well?
18:28:49 <cp16net> its abstract
18:29:01 <cp16net> thats what we focus on mainly
18:29:24 <cp16net> meaning the tests in the gate
18:30:40 <SlickNik> Seems reasonable to me. Any other feedback?
18:30:48 <amrith> cp16net, is there a way to list all templates?
18:31:06 <cp16net> yeah... flavor-list
18:31:50 <amrith> SlickNik, is there a decision we plan to take on this today?
18:33:06 <cp16net> it doesnt seem like it
18:33:16 <SlickNik> amrith: It looks good at a quick glance, but I'd like to give folks a bit more time to look at it before making a final decision.
18:33:24 <cp16net> sounds fine
18:33:28 <amrith> SlickNik, thanks
18:33:37 <amrith> I thought we could have a list of templates
18:33:41 <amrith> independent of flavors
18:33:46 <amrith> but I may be misunderstanding
18:33:52 <amrith> or the terminology may be tripping me up
18:34:08 <amrith> it looks like (from the example) that all the configuration is mysql specific
18:34:22 <amrith> and I'm wondering whether we need a custom flavor for each datastore (other than MySQL)
18:35:13 <nehav3> as an input you can pass a datastore and version id so you would get the params for other databases
18:35:15 <SlickNik> amrith: if you look at the GET resource, it's /datastore/version/{dv_id}/configuration/{flavor_id}
18:35:33 <SlickNik> so it depends on datastore-version, and flavor.
18:35:36 <nehav3> slicknick exactly
18:36:14 <SlickNik> the dependence on flavor is probably because your templates for your defaults could be written taking the flavor into account.
18:36:36 <nehav3> yes for optimizing the templates for each flavor
18:36:40 <amrith> hmm.. I see trove/templates/<datastore>/config.template files in the source tree
18:36:41 <SlickNik> if that's not the case, I would think the call would return the same defaults regardless of the flavor
18:36:45 <amcrn> nehav3 cp16net: generally looks good to me as well, but i'll take a stronger peek later today
18:37:00 <amrith> so I assume that this is the information that comes back
18:37:12 <cp16net> amrith: flex those muscles
18:37:47 <amrith> but that would mean that if I created a new template (non-default) where would it be visible (what's the URI)
18:38:09 <amrith> yes, I have to exercise some more ...
18:38:58 <SlickNik> amrith / cp16net / nehav: Can I suggest taking clarifications regarding this to after the meeting, since there were a couple more open items to cover?
18:39:07 <amrith> sure.
18:39:09 <nehav3> sure
18:39:22 <cp16net> +1
18:39:45 <SlickNik> Thanks — amrith, you had an item… go for it.
18:40:15 <amrith> I just wanted to know what denis_makogon was asking re: google calendar and stuff. I don't use google calendar and I don't know what it has to do with hangouts.
18:40:22 <amrith> so that can be after the meeting as well
18:40:29 <kevinconway> amrith: #link http://learn.googleapps.com/training-videos/cal_overview_vid
18:40:33 <amrith> kevinconway, may have valuable input though ;)
18:40:36 <amrith> there, he did!
18:40:39 <denis_makogon> amrith, i get it =)
18:40:44 <amrith> thanks, I'll watch the video
18:41:09 <SlickNik> okay, in that case. denis_makogon: you had an item for open discussion?
18:41:10 <kevinconway> amrith: it doesn't really have anything to do with hangouts. denis_makogon may be talking about setting up some scheduled google hangouts for the midcycle
18:41:40 <denis_makogon> kevinconway, for sure !
18:42:13 <amrith> ok denis_makogon what were you talking about? we're all trying to interpret your PM to me ... which I didn't understand, sorry
18:42:37 <denis_makogon> amrith, nevermind
18:42:50 <denis_makogon> there's another question, from me
18:43:01 <SlickNik> go for it denis_makogon
18:43:12 <denis_makogon> since replication v1 is almost finished
18:43:45 <denis_makogon> we might to write down a wiki page that describes a list of datastores and version that were used for testing it
18:44:02 <denis_makogon> some sort of "official datastores supported by Trove"
18:44:16 <kevinconway> would that be the feature matrix we had?
18:44:24 <dougshelley66> denis_makogon, i'm not clear why that has anything to do with replication. Wouldn't we want that for every feature?
18:44:28 <SlickNik> denis_makogon: Didn't we already have a feature matrix?
18:44:54 <cp16net> https://wiki.openstack.org/wiki/Trove/DatastoreCompatibilityMatrix
18:44:57 <cp16net> ?
18:45:00 <cp16net> #link https://wiki.openstack.org/wiki/Trove/DatastoreCompatibilityMatrix
18:45:03 <cp16net> there its linked now
18:45:04 <denis_makogon> SlickNik, yes we have, but there's no loud words like "official" etc.
18:45:50 <iccha> i dont think we want to use words like official, because anything which becomes officially supported by openstack i think should go through TC and must have gated tests
18:46:33 <grapex_> how about turnstile tests?
18:46:36 <iccha> although i do like adding it to wiki to keep track and notes of what we have tested
18:46:47 <denis_makogon> iccha, yes, agreed, but we still need something that states that Trove was well tested with
18:46:47 <SlickNik> iccha: +1
18:47:01 <kevinconway> grapex_: i like smoke tests
18:47:23 <grapex_> So...
18:47:37 <grapex_> I feel like this is a huge can of worms, so maybe we should just talk about talking about it at the summit
18:47:54 <grapex_> I just started looking through the cluster pull requests today
18:47:58 <dougshelley66> however, if the tests are automated and running regularly, it is unclear what value making any claims is
18:47:59 <grapex_> but it seems like its all just mongo, right?
18:48:05 <grapex_> so far anyway
18:48:13 <SlickNik> denis_makogon: So right now openstack gate tests exist only for mysql, and that's all the statements I think we can make.
18:48:17 <grapex_> but there are no mongo tests yet.
18:48:18 <dougshelley66> sorry, i mean aren't automated
18:48:38 <grapex_> dougshelley66: +1
18:48:54 <denis_makogon> SlickNik, if that's it, then ok
18:48:59 <dougshelley66> SlickNik, +1
18:49:39 <denis_makogon> but once we would have fine grained Tempest coverage, we would need to come back to this discussion
18:50:24 <grapex_> amcrn: Can the PRs as is be tested in fake mode?
18:50:25 <SlickNik> Once we have tests for other datastores, we can come back and revise this to — we test mysql + other_datastore in the gate.
18:50:38 <amcrn> grapex_: PRs?
18:50:49 <grapex_> The four or five Clustering pull requests
18:50:50 <grapex_> sorry
18:50:52 <cp16net> s/PRs/reviews/
18:50:56 <iccha> we should be adding other datastore tests a priorirty
18:50:57 <amcrn> ah, sorry.
18:51:03 <grapex_> amcrn: that was totally HDURAG of me
18:51:12 <amcrn> grapex_: i need to talk to you about that after this meeting :)
18:51:37 <amcrn> grapex_: in short, maybe?
18:51:53 <denis_makogon> iccha, they are there, in trove-integration
18:52:47 <denis_makogon> iccha, to have more tests for other datastores we need to implement all required features like backup/restore, at least for now
18:53:00 <denis_makogon> i guess that it
18:53:12 <denis_makogon> i've got answers to all my questions
18:53:16 <SlickNik> denis_makogon: That's not the same as being part of the gate. So I think the first step is to make them run as part of the gate tests.
18:53:31 <SlickNik> denis_makogon: even before implementing other features like backup / restore.
18:53:38 <iccha> could we add tests for all datastores and add skis of datastores which do not support that feature?
18:53:39 <denis_makogon> SlickNik, agreed
18:54:03 <iccha> yes first getting it in gate like SlickNik is priorirty
18:54:39 <denis_makogon> iccha, fist we need to be able to run dsvm tests properly
18:54:41 <grapex_> So gate== Tempest?
18:55:10 <SlickNik> grapex_: nope gate == tempest + trove-int-tests running on OpenStack CI
18:56:13 <SlickNik> I think it makes sense to not move _all_ our int tests to Tempest, so we'll still have those int-tests around.
18:56:17 <iccha> good point denis_makogon
18:56:26 <grapex_> SlickNik: Ok. I kind of wish instead of a design summit we could meet for a hackathon. :)
18:56:47 <grapex_> I feel like there's some underlying problems that if the cores paired up on we could fix if given a few days
18:56:48 <iccha> we could ditch the happy hpurs and baseball games and hack instead grapex_ :p
18:56:51 <amrith> grapex_, we could consider re-purposing some of the time for that ...
18:56:54 <grapex_> iccha: Surely you jest!
18:56:57 <SlickNik> grapex: maybe we can convert a couple of sessions to "work" hacking sessions.
18:57:03 <grapex_> amrith SlickNik: I'm down
18:57:39 <amrith> iccha, ... ditch happy hour ;(
18:57:48 <iccha> amrith: i was joking :)
18:57:50 <cp16net> i'm still going to a baseball game
18:57:53 <iccha> thats called team building
18:57:58 <SlickNik> Okay, let's talk about that.
18:57:58 <amrith> phew ...
18:58:01 <peterstac> iccha: We'd have to call it unhappy hour then :(
18:58:09 <SlickNik> Looks like that's all for today.
18:58:14 <SlickNik> #endmeeting