Wednesday, 2015-07-01

*** kfox1111 is now known as kfox1111_away00:33
*** kzaitsev_mb has joined #openstack-app-catalog04:32
*** kzaitsev_mb has quit IRC04:46
*** isviridov_away is now known as isviridov07:01
*** kzaitsev_mb has joined #openstack-app-catalog07:19
*** kzaitsev_mb has quit IRC07:33
*** kzaitsev_mb has joined #openstack-app-catalog07:34
*** kzaitsev_mb has quit IRC07:47
*** isviridov is now known as isviridov_away08:13
*** isviridov_away is now known as isviridov13:24
*** ativelkov has quit IRC14:46
*** ativelkov has joined #openstack-app-catalog14:47
*** kfox1111_away is now known as kfox111115:35
*** kzaitsev_mb has joined #openstack-app-catalog17:07
*** kzaitsev_mb has quit IRC17:18
*** isviridov is now known as isviridov_away17:56
*** reed_ has joined #openstack-app-catalog19:15
docaedoHello there app-catalogers! I've been on vaca this week so only doing a little bit of stuff, and haven't thought about the meeting tomorrow - anyone have anything specific for the agenda?19:39
docaedoI need to expand the commit message with more details around the why (https://review.openstack.org/194875)19:40
docaedoand am planning to take a first pass at combining the three YAMLs into a single one later today.19:40
docaedoLet me know if there are any topics for discussion anyone wants to cover tomorrow and I'll update the agenda and send an announcement.19:41
*** openstackgerrit has quit IRC20:37
*** openstackgerrit has joined #openstack-app-catalog20:37
kfox1111I've been discussing the heat templates a bit with the heat folks. we need to integrate env url support as well.22:39
kfox1111a status of the glance artifact poc would be good too.22:39
kfox1111docaedo: You could put that YAML isn't natively supported in the commit message, if you want to get more specific.22:42
kfox1111all modern web browsers parse json natively these days.22:42
docaedokfox1111: I'll see if that commit message is good enough for now, but will add that if the intention is still not clear23:03
docaedokfox1111: good news that you've had some discussion with the heat folks23:04
kfox1111kk.23:09
kfox1111yeah. was discussing the need for conditionals some more with them and they made an interesting statement about the template should be very generic and the env shoudl be where the specific stuff goes.23:09
kfox1111the thread wandered into, maybe the app catalog should provide multiple env's so the user could pick.23:10
kfox1111then it went back to, well, heat's going to support a feature where one env could be provided and heat will allow the user to select between the various implementation choices.23:10
kfox1111so they will eventually resolve the issue as part of heat, and we only need to support one env file.23:10
docaedoLike - have a generic template in app catalog, and then offer the user a way to choose what environment they want to support and the template is adjusted on the fly?23:11
kfox1111so that's good. we won't have to put any complicated engine logic in the app catalog.23:11
kfox1111yeah. something like that.23:11
kfox1111like, I want to write a generic template that supports both nova-network and neutron. thats kind of hard today.23:12
kfox1111also see the magnum heat template thread on the mailing list from a couple of days ago. :/23:12
docaedoGot it - yeah I would argue against that approach, wouldn't make sense for app catalog, would make more sense to have heat (or one of the template builder approaches) handle that23:12
* kfox1111 nods.23:12
kfox1111so if anyone expresses interest in that,  Itold the heat folks we would send the developers their way.23:13
docaedolike put things that work in the app catalog, and if things don't work, fix them first :)23:13
kfox1111we will just allow one env associated with the heat record, and let heat present the various options to the user.23:13
kfox1111the problem is, its increadibly difficult to make both a complicated cloud scaled application, and make it generic enough to contribute at the moment. :/23:14
kfox1111its a really big problem. :/23:14
kfox1111openstack cloud's are quite different from each other. :/23:15
kfox1111nova network vs neutron being one of the biggest offenders. :(23:15
docaedoto be fair, incredibly hard given the native tools available. Arguably this is easier if you're using mesos, CF, kubernetes (so you push that problem to a different abstraction layer)23:16
kfox1111no, cause those only deal with vms, not cloud resources like heat does.23:18
kfox1111example, designate dns as a service.23:18
kfox1111there is no easy way today to make optionally registering the vm with a dns entry as part of the deploy if its supported by the cloud.23:18
kfox1111either you add the feature to the heat template and it breaks if not available, or you dont. :/23:19
docaedoah true23:19
docaedoOK I'm going to be out for a while - if you want to add/remove anything from the agenda tomorrow, please doo :)23:27

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