17:01:51 <docaedo> #startmeeting app-catalog
17:01:51 <openstack> Meeting started Thu Jan 21 17:01:51 2016 UTC and is due to finish in 60 minutes.  The chair is docaedo. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:01:52 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:01:55 <openstack> The meeting name has been set to 'app_catalog'
17:02:01 <docaedo> courtest ping: kfox1111 spzala ativelkov kzaitsev_ws
17:02:11 <docaedo> #topic rollcall
17:02:11 <kzaitsev_ws> o/
17:02:14 <ativelkov> o/
17:02:18 <docaedo> o/
17:02:19 <spzala> o/
17:02:43 <docaedo> Our agenda is at
17:02:46 <docaedo> #link https://wiki.openstack.org/wiki/Meetings/app-catalog
17:03:05 <docaedo> I'm switching things up just a tiny bit so spzala can help gather up the info we'll need for TOSCA, so..
17:03:13 <docaedo> #topic Adding TOSCA to the catalog
17:03:23 <docaedo> First thing we need to cover is what gets hosted in the catalog?  My understanding is that we would have
17:03:26 <docaedo> CSAR (contains template, scripts, binaries etc) and individual templates - but I'll let spzala take over now :)
17:03:27 <spzala> thanks docaedo
17:04:15 <spzala> :-) OK, sure. As you said, CSAR is a container file and for now as support two ext .zip and .csar
17:04:30 <spzala> individual templates are .yaml
17:05:01 <spzala> so that's pretty much about what kinda format we can expect to host
17:05:09 <docaedo> I'm not as familiar as I would like to be when it comes to TOSCA - if I wanted to use a CSAR in my environment, would that be something I would have to unpack locally and feed to heat-translator or something? What's the use case like?
17:06:11 <spzala> yup, so for tosca-parser and heat-translator the CSAR is expected as a .zip or .csar input and then it gets unpacked locally on a temp directory
17:06:43 <spzala> and templates other artifacts in unpacked dir used for parsing and or tranlation
17:06:59 <spzala> so you feed csar as such
17:07:17 <spzala> and translator/parser will take it from there
17:07:22 <docaedo> ok so if I have an environment with tosca-parser installed, can I give it a URL for a CSAR file? (just wondering I guess about the client side of this)
17:07:32 <spzala> yup
17:07:57 <docaedo> ok cool, I think that makes sense to me
17:08:12 <spzala> OK cool
17:08:39 <spzala> and it's same for individual templates
17:09:00 <docaedo> so we need to identify the metadata that we'd want to see, to add to the assets schema .. seems like we're just talking about two asset types (csar and template), would both have the same metadata associated with them?
17:10:00 <spzala> I need to understand that from you now on .. is there any example out there with metadata that i can refer to
17:10:10 <spzala> for example how it looks like for Heat templates
17:11:10 <docaedo> sure - hold a sec for two quick example links
17:11:13 <spzala> I think we can keep it same, CSAR contains a meta file that translator/parser use to process it
17:11:21 <spzala> ok, thx
17:11:53 <docaedo> The schema is here:
17:11:56 <docaedo> #link https://git.openstack.org/cgit/openstack/app-catalog/tree/openstack_catalog/web/static/assets.schema.yaml
17:12:11 <docaedo> and all the current assets are here:
17:12:15 <docaedo> #link https://git.openstack.org/cgit/openstack/app-catalog/tree/openstack_catalog/web/static/assets.yaml
17:12:42 <docaedo> finally, example doc on the wiki here:
17:12:44 <docaedo> #link https://wiki.openstack.org/wiki/App-Catalog#Structure_of_Heat_entry
17:15:02 <spzala> thanks.. looking at it but need to understand it better something I would get back to you. Looking at the second link,
17:15:09 <docaedo> I think we'd want to do something similar to murano assets, where there's a "format" field
17:15:14 <spzala> the flow will be similar to Heat templates with little change
17:15:29 <spzala> yup I notice that at the end of the file
17:15:37 <docaedo> sure - the last link (from the wiki) probably is the cleanest view/example
17:16:08 <docaedo> and happy to have you thinking about it for a bit, and ask questions on #openstack-app-catalog as they come up
17:16:36 <docaedo> should be easy stuff :)  In the mean time I'll create a blueprint similar to the one I made for mistral
17:16:38 <spzala> yup it is.. sure, let me create a similar flow
17:16:39 <docaedo> #link https://blueprints.launchpad.net/app-catalog/+spec/add-mistral-assets
17:16:47 <docaedo> since the steps will be similar
17:16:48 <spzala> and I will send it for a review
17:16:58 <docaedo> good plan!
17:17:02 <spzala> OK, nice
17:17:47 <spzala> So that's an action item for me :-)
17:18:22 <docaedo> you got it - I'll make the blueprint right after this meeting (called add-tosca-assets)
17:18:28 <docaedo> and will share link on app catalog channel
17:18:41 <docaedo> any other questions/thoughts/issues on this topic?
17:19:17 <spzala> OK perfect
17:19:41 <spzala> nope, good plan as you said
17:19:49 <spzala> so no questions from me
17:19:57 <docaedo> ok, will move on to next topic momentarily (give anyone else a chance to wave)
17:20:22 <spzala> thanks docaedo and all.. I am stepping away.
17:20:43 <docaedo> spzala: no prob, thanks for joining today!
17:20:56 <docaedo> #topic Status updates
17:21:19 <spzala> doceado: no problem!
17:21:39 <docaedo> Yesterday kfox1111 pushed up some good stuff for making the catalog a little more "dynamic"
17:21:51 <docaedo> #link https://review.openstack.org/267087
17:22:17 <docaedo> that one is the key bit - it now determines the last modified time for each asset based on git
17:22:37 <docaedo> the last modified timestamp then shows up in the assets.json used by the website and the horizon plugin
17:23:08 <docaedo> first thing we'll do is work on a change to the site where it says "recently added apps" - that will become dynamic now, rather than static
17:23:22 <kzaitsev_mb> Would probably be nice to add option on site to sort by the field
17:23:34 <docaedo> yes that too (good reminder!)
17:23:47 <docaedo> I also changed the dead-link checker to work with this
17:23:49 <docaedo> #link https://review.openstack.org/269935
17:23:55 <kzaitsev_mb> recently added and/or recently updated =)
17:24:01 <docaedo> #link https://review.openstack.org/270579
17:24:54 <docaedo> exactly - and can probably even make it select the top 10 entries that fit (ie. last 10 modified entries) and do a random pick from those, so it rotates a bit on reload
17:25:07 <docaedo> otherwise it'll probably be mostly static anyway :)
17:25:37 <docaedo> reviews for 269935 would be good
17:26:33 <docaedo> any other status updates? Otherwise I think we can talk about Glare/Glance for a bit (and if anyone has additional topics, please speak up and we'll have open discussion too)
17:27:51 <docaedo> #topic Discuss Glare outline/glare implementation
17:27:58 <docaedo> #link https://etherpad.openstack.org/p/app-cat-glare
17:28:34 <docaedo> kzaitsev_mb: I think the outline is good, and the plan makes sense - and anyone who would have any strong objections has had a LOT of time to argue against
17:29:14 <docaedo> so I would like to start moving on this - have you been working on it? Are there any ways I can help out?
17:30:10 <kzaitsev_mb> I wanted to start working on it this Wednesday, but pip8 was released, and broke a couple of our gates =)
17:30:43 <kzaitsev_mb> So currently my closest plan would be to finish m-2 release chores and start makeing the glare plugin
17:30:51 <docaedo> haha yeah that pip 8 fun, very exciting
17:31:27 <docaedo> that's great - please put up status updates occasionally on the channel, especially if there's a WIP I could deploy locally to mess around with
17:31:39 <kzaitsev_mb> now we have pep8 which is good and pip8 that isn't (=
17:31:42 <docaedo> For the production implementation assume we're using trove for the DB btw
17:32:30 <kzaitsev_mb> docaedo: would surely do, so.
17:32:49 <docaedo> kzaitsev_mb: thanks!
17:33:10 <docaedo> I think we're good then, and maybe done?  Anyone else here want to talk about anything?
17:33:22 * docaedo waits one full minute at least
17:33:50 <kzaitsev_mb> does trove act as a drop-in server replacement? just wondering, haven't played with it yet
17:35:27 <docaedo> I *believe* so, you basically get a mysql server, you just don't have to manage it (because infra isn't going to deal with managing a mysql server on the app catalog host - AFAIK they don't do that for anyway, just have everyone use trove)
17:36:00 <kzaitsev_mb> we'll find out then =)
17:36:08 <docaedo> I think for testing, whatever the minimal local thing glare needs is fine, and when we get closer to something for production we can work out the details, but shouldn't require anything other than basically planning to use a mysql server
17:36:25 <docaedo> oh we can also assume we'll have access to swift for object storage
17:36:41 <docaedo> (which of course makes the footprint for testing this .. non-trivial?)
17:37:12 <kzaitsev_mb> and sounds fair about infra folks, so yeah, using trove makes sense then
17:38:27 <docaedo> any questions or issues that you hit though, ask on the app-catalog channel and I'll figure it out
17:39:26 <docaedo> seems like we're done for today? I don't have anything else, and I believe it's just kzaitsev_mb and me at this point :)
17:40:07 <kzaitsev_mb> yep, seems, that we're done. Nice to see more assets flowing into the catalog =)
17:40:35 <docaedo> yeah agreed - will be great to get mistral and tosca stuff in there
17:40:52 <docaedo> well thanks for coming today, looking very much forward to seeing/helping with glare stuff!
17:41:22 <docaedo> #endmeeting