Tuesday, 2015-12-15

*** kzaitsev_mb has joined #openstack-app-catalog00:25
*** reed_ has quit IRC00:59
*** kebray has quit IRC01:01
*** openstack has joined #openstack-app-catalog01:09
*** kzaitsev_mb has quit IRC01:29
openstackgerritChristopher Aedo proposed openstack/app-catalog: Skip assets with active==false attribute  https://review.openstack.org/25766303:12
*** kebray has joined #openstack-app-catalog04:12
*** kebray has quit IRC04:14
*** kebray has joined #openstack-app-catalog04:14
*** openstack has joined #openstack-app-catalog04:27
docaedo^^ first step for dead link checker (and hash updater). Just adjusts the website to let attributes.active=false hide an asset.05:01
*** olaph has quit IRC05:07
*** kebray has quit IRC07:19
*** openstackgerrit has quit IRC09:47
*** openstackgerrit has joined #openstack-app-catalog09:48
*** openstackgerrit has quit IRC10:32
*** openstackgerrit has joined #openstack-app-catalog10:33
*** kzaitsev_mb has joined #openstack-app-catalog10:34
*** ig0r_ has quit IRC11:20
*** dslevin has quit IRC14:42
*** dslevin has joined #openstack-app-catalog15:06
*** kzaitsev_mb has quit IRC15:15
*** kzaitsev_mb has joined #openstack-app-catalog15:16
*** dslevin has quit IRC16:06
*** kzaitsev_mb has quit IRC16:12
*** kebray has joined #openstack-app-catalog16:56
*** kzaitsev_mb has joined #openstack-app-catalog17:05
kfox1111kzaitsev_mb: which etherpad?17:21
kfox1111docaedo: we should probably do two things. one, move the filtering code to the python api. that way, it will work with both the website and the horizion plugin.17:22
kfox1111two, put the active: false things in a second document, that we merge in. that way, we don't have to check it in. the link checker can just update it.17:23
kfox1111we'll probaly need something like that anyway for stars and reviews and such. so starting now wouldn't hurt.17:23
kzaitsev_mbkfox1111: https://etherpad.openstack.org/p/app-cat-glare this one17:24
kzaitsev_mbI've been planning on getting my hands on the glare backend around next week, hopefully =)17:24
docaedoI am heavily against making stars and rating work in some crazy mashup with the current assets.yaml17:26
docaedoAs soon as we have an API that handles authentication, we need to switch to that being the only way to add/update/remove content17:27
kfox1111just a union for now.17:27
docaedobut we can't have stars and ratings without authentication17:28
kfox1111hmm... true.17:28
kfox1111I think we're still kind of far from having glare back stuff though,17:28
docaedothanks by the way for pointing out the javascript "fix" should have been in python :) I wish I'd thought of that before I poked around in JS17:28
kfox1111so some kind of intermediate step would be good so we can make some visiable progress.17:29
docaedoI'd be up for thinking/talking about something we can do in the mean time, but for me at least the next thing that I can get done soon (maybe even this week) is an update-bot that kills dead links, and keeps some entries updated with a new hash (and an "updated" date field in attributes)17:32
*** kebray has quit IRC17:32
kfox1111yeah. the update bot would be awesome.17:33
docaedoupdating yaml2json.py to exclude attributes.active==false will be easy, and gets those entries pulled out for both web site and horizon right?17:33
kfox1111still, the dead links should be kept seperate from assets.yaml.17:33
kfox1111yeah. yaml2json.py could load in assets.yaml and linkchecker.yaml or something, and then filter out the dead stuff.17:33
docaedowhy? If they're in assets.yaml, then the "owner" of that asset can easily repair their link and drop that active:false17:34
kfox1111because when someone updates the assets.yaml,17:34
kfox1111all the dead links suddenly go alive again, until the next dead link check pass.17:34
kfox1111the links should probably stay dead until the link checker revalidates them.17:35
kfox1111the updating of the assets.yaml file could kick off another link checker pass though.17:35
docaedono, the dead link bot submits a gerrit review that still has to be merged - if we suddenly had multiple submissions per day that would be a concern, but I don't see that being something to worry about right now?17:35
kfox1111I guess that could work too, but bots submitting gerrit reviews like that just feels......17:36
kfox1111dirty. :/17:36
docaedoI'm using the same mechanism infra uses for normalizing project requirements :) agree it's not a great choice for many cases, but in this one I think it makes the most sense until we get out of updating that yaml file17:38
kfox1111why would merging two yaml files, one maintained by gerrit, one by the link checker be bad?17:38
*** kebray has joined #openstack-app-catalog17:38
kfox1111that lets the authorative source of each of the data own it.17:38
kfox1111then users don't have to care about active at all.17:39
kfox1111(nor we, for that matter)17:39
kfox1111kzaitsev_mb: This is a good doc, but long. I don't have time to fully review it today. please keep bugging me about it.17:39
kzaitsev_mbkfox1111: would there be a way for a living person to speed up a live resource becoming active? Or would he only be restricted for having to wait for rechecker?17:40
docaedoTo me the merging idea seems to add unnecessary complication, and makes things a whole lot less obvious to newcomers17:40
kzaitsev_mbkfox1111: sure, will do? )17:40
docaedobut also, nobody has to care about active unless their asset is marked active:false17:41
kfox1111got a meeting. be back in a bit. :/17:43
docaedoenjoy it!17:43
kfox1111the merging is like one extra python line I think.17:43
docaedotrue - and we can debate more when you get back from the meeting - but if the bot is updating a different YAML that is then merged on the server before being turned into a JSON, I still don't really see the advantage?17:44
docaedoand if you're suggesting the bot not submit via gerrit, and just have the server merge whatever doc the bot comes up with, well that scares me a lot :) but I don't think that's what you're suggesting17:45
kzaitsev_mbdocaedo: have you had the chance to update the apps?17:48
docaedokzaitsev_mb: yes, updated - sorry I thought I said yesterday but maybe I didn't.  OH I know, I updated them then +2'd the review17:49
docaedoanyway they're updated, we can merge the assets.yaml now17:49
kzaitsev_mbthanks! )17:50
docaedolast point re: "why would merging two yaml files, one maintained by gerrit, one by the link checker be bad?" .. I fear a case of the link checker having a terrible DNS failure or something when it runs one time, and invalidating every asset in the catalog - how do we undo that run? vs. every change having human input, easy to just NOT merge that run17:52
kzaitsev_mbhm. wierd. when I download one it still seems to have the old manifest17:52
kzaitsev_mblet me recheck17:53
docaedoodds are this bot will almost never actually submit a review with updated dead links with the current size of the catalog - and it becomes a different issue as the catalog gets much much bigger (but presumably we have different problems and different mechanisms by then)17:53
docaedokzaitsev_mb: which one? Maybe I need to flush the whole CDN just to be sae17:53
docaedohahaha  SAFE17:53
kzaitsev_mbI'm checking this one http://apps.openstack.org/#tab=murano-apps&asset=Docker%20HTTPd it should have Interfaces in Requires line in manifest17:54
kzaitsev_mblet me check the apps I sent you, I might have packaged them wrong17:55
kzaitsev_mbwell at least locally they seem to be correct17:55
kzaitsev_mbis there any way to flush the CDN? sounds like that might be the thing that needs to be done17:57
docaedoyep, I'm on it17:58
docaedofunny, I pulled down (from storage.apps.o.o) all the files I uploaded yesterday, seems like about half have different hashs - so will trigger a cache flush as needed18:05
docaedoOh rackspace! "Only 25 cdn object purges allowed per day."18:12
docaedoso purged what I could, I think we're going to be fine though - maybe I'm overly pessimistic but I don't think there will be TOO many people impacted by getting the old version of the .zip until the cache expires (within 24 hours of the upload)18:13
*** kzaitsev_mb has quit IRC18:13
*** kzaitsev_mb has joined #openstack-app-catalog18:17
kzaitsev_mbseems that did the trick18:17
kzaitsev_mbthanks again!18:17
openstackgerritMerged openstack/app-catalog: Update murano apps to be forward compatible with liberty  https://review.openstack.org/25747618:19
docaedono prob18:22
docaedoalso, this merge has me holding my breath and watching puppetboard18:22
docaedoThis finally pushed the updated local_settings which turns off debug18:24
docaedobut as far as I can tell, didn't break anything else18:24
*** kzaitsev_mb has quit IRC18:26
docaedoalso did some other stuff - this was all part of the last infra change I put in that sorted out the ordering properly18:27
*** kzaitsev_mb has joined #openstack-app-catalog18:27
* docaedo breathes a sigh of relief18:29
*** kzaitsev_mb has quit IRC19:19
docaedoI'm curious/confused about this entry: http://apps.openstack.org/#tab=heat-templates&asset=Hello%20World19:28
docaedoSpecifically this part:19:28
docaedo        environment:19:28
docaedo          parameters:19:28
docaedo            db_port: 100019:28
docaedoWhy would the app catalog entry for a heat template include any parameters like that?19:29
docaedoIt's the only heat entry with that in it, and I suspect it's a mistake - I'm inclined to replace that entry with the canonical hello world heat template in the heat examples repo19:31
*** openstackgerrit has quit IRC19:32
*** openstackgerrit has joined #openstack-app-catalog19:33
kfox1111docaedo: it was a test of the environment schema. doesn't matter. just showing how it could work.19:38
*** kzaitsev_mb has joined #openstack-app-catalog19:38
docaedook - I guess I'm confused about the environment schema then, not sure what it's meant to accomplish?19:38
kfox1111the hello world template's a bit of a silly thing to do for an app catalog anyway. no one's ever going to use it.19:38
kfox1111anything under environment: follows the heat environment schema.19:39
kfox1111  db_port: 100019:39
docaedoI noticed it because I was looking at the yaml2json script19:39
kfox1111is a valid heat env file.19:39
docaedooh I get it - so if the environment details are in the asset entry, someone could use the horizon plugin to launch the heat template without having to do anything?19:40
docaedowell that makes sense, thanks :)19:41
kfox1111np. :)19:41
kfox1111I'll read back in the scrollback buffer later. still have 3 pressing things all at once going on. :/19:42
docaedonp, seems like you have a lot going on so don't sweat it19:42
docaedoI'll propose something and we can argue at our leisure via gerrit, like all the other projects do ;)19:42
*** kebray has quit IRC19:45
*** kzaitsev_mb has quit IRC19:59
*** kzaitsev_mb has joined #openstack-app-catalog20:05
kfox1111arg... another example where heat 2015.2.0 through 2015.2.1 fails with a template, but 2015.2.2 contains the bugfix needed...20:46
openstackLaunchpad bug 1473660 in heat "multipart mime wait handle" [High,Fix released] - Assigned to Rabi Mishra (rabi)20:47
*** dslevin has joined #openstack-app-catalog20:47
*** kzaitsev_mb has quit IRC20:58
openstackgerritChristopher Aedo proposed openstack/app-catalog: Skip assets with active==false attribute  https://review.openstack.org/25766322:26
openstackgerritChristopher Aedo proposed openstack/app-catalog: Skip assets with active==false attribute  https://review.openstack.org/25766322:34
*** kzaitsev_mb has joined #openstack-app-catalog23:31

Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!