17:00:31 <docaedo> #startmeeting app-catalog
17:00:31 <openstack> Meeting started Thu Dec 10 17:00:31 2015 UTC and is due to finish in 60 minutes.  The chair is docaedo. Information about MeetBot at http://wiki.debian.org/MeetBot.
17:00:33 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
17:00:36 <openstack> The meeting name has been set to 'app_catalog'
17:00:58 <docaedo> #topic rollcall
17:01:24 <docaedo> o/
17:01:25 <kzaitsev_mb> o/
17:02:22 <docaedo> courtesy ping for kfox1111 olaph
17:02:33 * olaph waves
17:02:48 <docaedo> The agenda for today is light :) https://wiki.openstack.org/wiki/Meetings/app-catalog
17:03:15 <docaedo> #topic Status updates
17:04:04 <docaedo> I don't actually have any (have been off-site this week so have done not too much related to app-catalog) .. kzaitsev_mb, any updates from your world?
17:04:47 <kzaitsev_mb> not really, I've ben fixing our CI/gate all week. =(
17:04:58 <docaedo> ah that sounds fun!
17:05:18 <docaedo> ok moving on :)
17:05:29 <kzaitsev_mb> One interesting thing I heard was that nikhil from glance wanted to make a conference call about artifacts. Should look into that
17:05:45 <docaedo> #topic Discuss Glare outline
17:05:49 <docaedo> #link https://etherpad.openstack.org/p/app-cat-glare
17:06:08 <docaedo> I'd be really curious to hear about a conference call? Why not an IRC meeting? Or is it just an internal thing?
17:06:31 <docaedo> BTW the glance artifact weekly meeting happens too early for me to attend, otherwise I'd be there
17:06:50 <kzaitsev_mb> I'm actually not entirelly sure — it was mentioned on todays glance meeting (drivers I suppose)
17:07:01 <kzaitsev_mb> I actually didn't attend, ativelkov pinged me on that one )
17:07:10 <kzaitsev_mb> I'll look into that and post some links =)
17:07:36 <docaedo> got it - well if there's a meeting coming, yes please share on #openstack-app-catalog, as I'll attend if at all possible
17:07:51 <docaedo> So on topic - I added just a few things to the etherpad, but generally agree with the plan
17:09:03 <kzaitsev_mb> cool, I thought that it can be an iterative and non-disrupting transition and tried to align it that way
17:09:35 <docaedo> yep, since the endpoint will be api/v2 anyway, we can do this without impacting anything current
17:10:22 <docaedo> Otherwise I basically am on board with all of it, just looking forward to the next steps for a PoC we can play with
17:11:37 <docaedo> Anything else on glare to share?
17:11:57 <kzaitsev_mb> Just read a comment of yours. So do you suggest just moving to option 2 right away, without the intermediate process?
17:12:29 <kzaitsev_mb> i.e makeing the web form and stuff and abandoning the gerrit-based approach asap? =)
17:12:45 <docaedo> yes basically
17:13:18 <kzaitsev_mb> ok. I was a bit worried, that it would probably take bit more time, since it implies the approval interface
17:13:27 <docaedo> I think it's going to be a while before we can abandon it, but when authentication works for that, and the glare bits are doing what we expect them to do, I don't see any reason to keep pushing changes through via updating a yaml
17:13:57 <docaedo> I think it's definitely going to take time - unless there's more work happening in the background somewhere, I haven't seen anything since tokyo
17:14:47 <docaedo> I would be happily impressed if we have something functional in two months :)
17:15:06 <kzaitsev_mb> sounds fair =)
17:16:12 <docaedo> At this point as far as I can tell, any further progress on that front hinges on you and ativelkov, so just keep us up to date
17:17:22 <docaedo> so that brings us I guess to the next topic? will hold for kzaitsev_mb to approve moving on...
17:17:30 <kzaitsev_mb> Sure. I believe, that I'll be able to dedicate most of my time to the task starting th week after next one.
17:17:40 <kzaitsev_mb> I think it would be holidays for you though =)
17:18:27 <kzaitsev_mb> well, nothing from my side, happy to hear that you approve the direction )
17:18:43 <docaedo> kzaitsev_mb: that's great news! will be holidays, but because OpenStack never sleeps, I will be around
17:19:02 <docaedo> so look forward to chatting about it on the IRC channel
17:19:15 <docaedo> #topic Dead link check/asset hash update script
17:19:54 <docaedo> I was talking with friends about intermediate approaches to keeping asset records updated in the catalog before there's a full API for it (i.e. glare implementation)
17:20:17 <docaedo> The suggestion I thought was pretty good, and I plan to work on this next week - basic outline is
17:20:44 <docaedo> a record can add an addition field under attributes, saying "auto-update: true"
17:21:28 <docaedo> if so, that record also have to have a field for a place to find the hash the asset SHOULD have
17:21:38 <docaedo> (use case being something like weekly builds of debian)
17:21:47 <kzaitsev_mb> is it a good idea to treat absence of this parameter as true?
17:21:59 <kzaitsev_mb> I mean what's the use case for "auto-update: false"?
17:22:28 <docaedo> generally auto-update: false would be "leave this record as-is"
17:22:32 <docaedo> so the rest of this:
17:22:41 <kzaitsev_mb> ah, I see what you mean
17:23:12 <docaedo> a daily script run looks at all the records that are told to auto-update, then compares expected hash against the hash listed in the catalog
17:23:27 <docaedo> if it's different, it updates the hash in the catalog, and submits that as a regular review
17:24:13 <docaedo> The same script will check ALL assets for dead links, and will <do something> with them (probably add a field like "active: false")
17:24:49 <kzaitsev_mb> now I get it =) it's not for the stored assets.
17:24:56 <kzaitsev_mb> makes total sense, though
17:25:28 <docaedo> kzaitsev_mb: yes, exactly - stored assets are a whole different problem that I think are not worth dealing with until we have glare in place
17:27:34 <docaedo> I'll provide updates on this and hope to have the first pass of the script that does this ready next week, should be reasonably straighforward
17:28:28 <docaedo> moving on
17:28:37 <docaedo> #topic Open discussion
17:29:48 <docaedo> Any open discussions to be had?
17:30:49 <kzaitsev_mb> Oh, just have one small thing to say, I'm being pinged by my colleagues who want me to update a couple of murano apps on the catalog.
17:30:50 <docaedo> Oh I also owe a followup re: TOSCA, but have not yet circled back with the folks I was talking to last week
17:31:27 <kzaitsev_mb> They've been doing it for some time already and I would probably get the binaries ready tomorrow and would ask you to upload them if that's ok
17:31:34 <docaedo> kzaitsev_mb: go for it - or do you mean update the binary/zip? if so, we can coordinate on the app-catalog channel and I can upload them
17:31:41 <docaedo> kzaitsev_mb: yep yep works great
17:32:03 <kzaitsev_mb> I've been trying to hold off the urge to update the asset after every commit )
17:33:04 <docaedo> oh the hash will change (I think the murano assets include hash right?) if so, just have to coordinate the update and upload
17:33:52 <kzaitsev_mb> ok, great, planning to ping you tomorrow with everything ready for that )
17:34:22 <docaedo> excellent!
17:34:40 <docaedo> well .. I'll wait two minutes to see if there's anything else, otherwise I think we are done
17:35:20 <nibalizer> is there free hosting for images somewhere? or do I need to host images myself?
17:35:46 <docaedo> Yes, free hosting is via the RAX cloud files we have associated with the catalog
17:36:07 <docaedo> the wiki has the instructions, basically source URL and hash need to be specified in the commit
17:36:31 <docaedo> then we grab it and put it up on cloud files before merging the update to assets.yaml. It's very very suboptimal
17:36:44 <nibalizer> awesome thansk
17:37:00 <docaedo> and will go away when glare is hosting the bits (there will still be something like coud files or swift behind it, but will be automagical then)
17:38:00 <docaedo> Think we are done then - thanks!
17:38:15 <docaedo> #endmeeting