| *** kgriffs is now known as kgriffs_afk | 00:00 | |
| *** blomquisg has joined #heat | 00:02 | |
| *** Trozz has joined #heat | 00:03 | |
| *** duncanjw has quit IRC | 00:03 | |
| *** pvaneck has quit IRC | 00:05 | |
| *** DaveJ__ has quit IRC | 00:17 | |
| *** mestery has quit IRC | 00:23 | |
| openstackgerrit | A change was merged to openstack/heat: Change router's prop name agent_id to l3_agent_id https://review.openstack.org/82716 | 00:25 |
|---|---|---|
| *** matsuhashi has joined #heat | 00:27 | |
| *** alexheneveld has quit IRC | 00:32 | |
| *** duncanjw has joined #heat | 00:33 | |
| *** blomquisg has quit IRC | 00:38 | |
| *** lindsayk has quit IRC | 00:41 | |
| *** gokrokve has joined #heat | 00:42 | |
| *** alexheneveld has joined #heat | 00:43 | |
| *** kfox1111 has quit IRC | 00:44 | |
| *** blomquisg has joined #heat | 00:51 | |
| *** duncanjw has quit IRC | 01:03 | |
| openstackgerrit | A change was merged to openstack/heat: Make OS::Nova::Server networks property updatable https://review.openstack.org/74299 | 01:05 |
| *** aweiteka has joined #heat | 01:06 | |
| *** matsuhashi has quit IRC | 01:12 | |
| *** matsuhashi has joined #heat | 01:14 | |
| *** sdake_ has quit IRC | 01:15 | |
| *** sdake has quit IRC | 01:16 | |
| *** networkn8 has quit IRC | 01:16 | |
| *** nosnos has joined #heat | 01:27 | |
| *** ramishra has joined #heat | 01:30 | |
| *** sdake has joined #heat | 01:35 | |
| *** sdake_ has joined #heat | 01:35 | |
| *** Linz has quit IRC | 01:37 | |
| *** duncanjw has joined #heat | 01:44 | |
| *** blomquisg has quit IRC | 01:45 | |
| *** alexpilotti has quit IRC | 01:47 | |
| *** zaneb has quit IRC | 01:48 | |
| *** zaneb has joined #heat | 01:49 | |
| *** mkollaro has quit IRC | 01:51 | |
| *** matsuhashi has quit IRC | 02:00 | |
| *** julienvey has joined #heat | 02:04 | |
| *** matsuhashi has joined #heat | 02:08 | |
| *** duncanjw has quit IRC | 02:20 | |
| *** alexheneveld has quit IRC | 02:25 | |
| *** harlowja is now known as harlowja_away | 02:27 | |
| *** alexheneveld has joined #heat | 02:27 | |
| *** DandyPandy_ has quit IRC | 02:27 | |
| *** aweiteka has quit IRC | 02:30 | |
| *** saurabhs1 has quit IRC | 02:32 | |
| *** DandyPandy has joined #heat | 02:33 | |
| *** fandi has quit IRC | 02:34 | |
| *** lazy_prince has quit IRC | 02:39 | |
| *** killer_prince has joined #heat | 02:39 | |
| *** nati_ueno has quit IRC | 02:42 | |
| *** julienvey has quit IRC | 02:45 | |
| *** ramishra has quit IRC | 02:45 | |
| *** ramishra has joined #heat | 02:45 | |
| *** ramishra has quit IRC | 02:50 | |
| *** alexheneveld has quit IRC | 02:53 | |
| *** ramishra has joined #heat | 02:54 | |
| *** alexheneveld has joined #heat | 02:55 | |
| *** alexheneveld_ has joined #heat | 02:57 | |
| *** alexheneveld has quit IRC | 02:59 | |
| *** alexheneveld_ is now known as alexheneveld | 02:59 | |
| openstackgerrit | Rabi Mishra proposed a change to openstack/heat: Provides ability to reference existing neutron resources https://review.openstack.org/82853 | 03:01 |
| *** ramishra has quit IRC | 03:02 | |
| *** ramishra has joined #heat | 03:02 | |
| *** ramishra has quit IRC | 03:07 | |
| *** zns has quit IRC | 03:12 | |
| *** qinjin has joined #heat | 03:13 | |
| *** matsuhashi has quit IRC | 03:15 | |
| *** nosnos has quit IRC | 03:28 | |
| *** samstav has quit IRC | 03:34 | |
| *** gokrokve has quit IRC | 03:37 | |
| *** gokrokve_ has joined #heat | 03:38 | |
| *** gokrokve_ has quit IRC | 03:42 | |
| *** wchrisj has quit IRC | 03:43 | |
| *** sgordon has joined #heat | 03:44 | |
| *** ramishra has joined #heat | 03:46 | |
| *** blamar has quit IRC | 03:46 | |
| *** mestery has joined #heat | 03:51 | |
| *** killer_prince is now known as lazy_prince | 03:53 | |
| *** blamar has joined #heat | 03:56 | |
| *** blamar has quit IRC | 03:57 | |
| *** tango has joined #heat | 03:58 | |
| *** blamar has joined #heat | 03:58 | |
| *** nati_ueno has joined #heat | 04:02 | |
| *** gokrokve has joined #heat | 04:07 | |
| *** blamar has quit IRC | 04:10 | |
| *** gokrokve has quit IRC | 04:13 | |
| *** nkhare has joined #heat | 04:14 | |
| *** david-lyle has joined #heat | 04:16 | |
| *** sgordon has quit IRC | 04:20 | |
| *** duncanjw has joined #heat | 04:21 | |
| *** pablosan has quit IRC | 04:22 | |
| *** pablosan has joined #heat | 04:22 | |
| *** matsuhashi has joined #heat | 04:25 | |
| *** mestery has quit IRC | 04:25 | |
| *** duncanjw has quit IRC | 04:25 | |
| *** cmyster has joined #heat | 04:26 | |
| cmyster | morning | 04:27 |
| *** wwallnrr__ has quit IRC | 04:30 | |
| *** nosnos has joined #heat | 04:36 | |
| *** lipinski1 has quit IRC | 04:38 | |
| *** blamar has joined #heat | 04:46 | |
| *** cfriesen has joined #heat | 05:03 | |
| *** ramishra_ has joined #heat | 05:04 | |
| *** ramishra has quit IRC | 05:07 | |
| *** skraynev_afk is now known as skraynev | 05:22 | |
| skraynev | Morning | 05:22 |
| openstackgerrit | Rabi Mishra proposed a change to openstack/heat: Provides ability to reference existing neutron resources https://review.openstack.org/82853 | 05:28 |
| *** lazy_prince has quit IRC | 05:28 | |
| *** nanjj has joined #heat | 05:34 | |
| openstackgerrit | Jun Jie Nan proposed a change to openstack/heat: Adjust user_data_format for server with deployment https://review.openstack.org/82992 | 05:44 |
| *** fandi has joined #heat | 05:44 | |
| nanjj | stevebaker: I tried several other approaches, only the one you mentioned works | 05:46 |
| *** e0ne has joined #heat | 05:55 | |
| *** tango has quit IRC | 05:57 | |
| *** nati_ueno has quit IRC | 05:57 | |
| *** lazy_prince has joined #heat | 06:00 | |
| *** IlyaE has joined #heat | 06:03 | |
| *** gokrokve has joined #heat | 06:08 | |
| openstackgerrit | Jenkins proposed a change to openstack/heat: Imported Translations from Transifex https://review.openstack.org/72566 | 06:09 |
| *** alexheneveld_ has joined #heat | 06:10 | |
| *** sdake has quit IRC | 06:12 | |
| *** alexheneveld has quit IRC | 06:12 | |
| *** alexheneveld_ is now known as alexheneveld | 06:12 | |
| *** sdake_ has quit IRC | 06:12 | |
| *** bvandenh has quit IRC | 06:13 | |
| *** gokrokve has quit IRC | 06:13 | |
| *** yogesh_ has joined #heat | 06:18 | |
| *** IlyaE has quit IRC | 06:18 | |
| *** saju_m has joined #heat | 06:19 | |
| *** duncanjw has joined #heat | 06:22 | |
| *** sdake has joined #heat | 06:24 | |
| *** sdake has joined #heat | 06:24 | |
| *** Darnoth has quit IRC | 06:25 | |
| *** duncanjw has quit IRC | 06:26 | |
| *** tnurlygayanov_ has joined #heat | 06:27 | |
| *** sdake_ has joined #heat | 06:27 | |
| *** tnurlygayanov has quit IRC | 06:27 | |
| *** alexheneveld_ has joined #heat | 06:29 | |
| *** alexheneveld has quit IRC | 06:31 | |
| *** alexheneveld_ is now known as alexheneveld | 06:31 | |
| openstackgerrit | Sergey Kraynev proposed a change to openstack/heat: Initial validation of functions https://review.openstack.org/82486 | 06:32 |
| *** duncanjw has joined #heat | 06:32 | |
| openstackgerrit | Sergey Kraynev proposed a change to openstack/heat: Validation functions for resources and outputs https://review.openstack.org/82487 | 06:32 |
| *** matsuhashi has quit IRC | 06:33 | |
| *** matsuhashi has joined #heat | 06:35 | |
| *** duncanjw has quit IRC | 06:37 | |
| *** cfriesen has quit IRC | 06:38 | |
| *** chen has joined #heat | 06:40 | |
| chen | hello, anyone here ?? I'm a new user to heat, just installed one on CentOS. I copied template file from https://github.com/openstack/heat-templates/blob/master/hot/servers_in_existing_neutron_network_no_floating_ips.yaml, and try to run it. But I get error :"ERROR: Unknown resource Type : OS::Neutron::SecurityGroup" | 06:42 |
| chen | Anyone know why this happen ?? | 06:42 |
| *** IlyaE has joined #heat | 06:43 | |
| openstackgerrit | Sergey Kraynev proposed a change to openstack/heat: Validation functions for resources and outputs https://review.openstack.org/82487 | 06:45 |
| *** e0ne has quit IRC | 06:47 | |
| *** e0ne has joined #heat | 06:48 | |
| *** e0ne has quit IRC | 06:51 | |
| openstackgerrit | huangtianhua proposed a change to openstack/heat: Make sure NetworkInterfaces and SubnetId updatable https://review.openstack.org/80209 | 06:56 |
| *** IlyaE has quit IRC | 07:03 | |
| *** matsuhas_ has joined #heat | 07:03 | |
| *** matsuhas_ has quit IRC | 07:04 | |
| *** alexheneveld has quit IRC | 07:05 | |
| *** matsuhas_ has joined #heat | 07:06 | |
| *** tspatzier has joined #heat | 07:06 | |
| *** matsuhashi has quit IRC | 07:07 | |
| *** saju_m has quit IRC | 07:13 | |
| openstackgerrit | Rabi Mishra proposed a change to openstack/heat: Provides ability to reference existing neutron resources https://review.openstack.org/82853 | 07:18 |
| *** yogesh_ has quit IRC | 07:19 | |
| *** yogesh_ has joined #heat | 07:19 | |
| *** ramishra_ has quit IRC | 07:34 | |
| *** jprovazn has joined #heat | 07:39 | |
| *** matsuhas_ has quit IRC | 07:59 | |
| openstackgerrit | Jun Jie Nan proposed a change to openstack/heat-templates: Software config hook for configuring with Chef https://review.openstack.org/80229 | 08:02 |
| openstackgerrit | Jun Jie Nan proposed a change to openstack/heat-templates: Add one example to show deploy sequence https://review.openstack.org/81757 | 08:02 |
| openstackgerrit | Jun Jie Nan proposed a change to openstack/heat-templates: Software config hook to provision using shell scripts https://review.openstack.org/70297 | 08:02 |
| openstackgerrit | Jun Jie Nan proposed a change to openstack/heat-templates: Software config hook to provision using puppet https://review.openstack.org/70639 | 08:02 |
| openstackgerrit | Jun Jie Nan proposed a change to openstack/heat-templates: Software config hook to provision using cfn-init https://review.openstack.org/79759 | 08:02 |
| openstackgerrit | Jun Jie Nan proposed a change to openstack/heat-templates: os-refresh-config script to invoke config hooks https://review.openstack.org/79758 | 08:02 |
| *** matsuhashi has joined #heat | 08:04 | |
| *** bvandenh has joined #heat | 08:08 | |
| *** gokrokve has joined #heat | 08:08 | |
| *** gokrokve has quit IRC | 08:12 | |
| cmyster | chen: what version ? | 08:15 |
| cmyster | Hi BTW | 08:15 |
| shardy | morning all | 08:15 |
| cmyster | morning shardy | 08:15 |
| *** tomek_adamczewsk has joined #heat | 08:15 | |
| therve | Good morning! | 08:18 |
| chen | cmyster Havana | 08:21 |
| cmyster | checn take a look here http://docs.openstack.org/developer/heat/template_guide/openstack.html#OS::Neutron::SecurityGroup seems like it *is* there... | 08:21 |
| cmyster | chen: pastebin yer template please ? | 08:21 |
| *** jstrachan has joined #heat | 08:21 | |
| cmyster | morning therve | 08:21 |
| chen | cmyster http://paste.openstack.org/show/74319/ | 08:23 |
| chen | cmyster I directly copied it from github | 08:23 |
| *** yogesh_ has quit IRC | 08:25 | |
| *** pas-ha has quit IRC | 08:26 | |
| therve | chen, What's your heat version is the question I think | 08:27 |
| cmyster | chen: strange, cazn you please paste the command you tried ? | 08:27 |
| cmyster | with all the parameters | 08:27 |
| chen | therve I installed it using "yum install...." | 08:27 |
| therve | chen, I'm sure you can get some information out of yum then | 08:28 |
| skraynev | chen, cmyster: I suppose that OS::Neutron::SecurityGroup just not available in havana release | 08:28 |
| chen | heat stack-create mystack --template-file=servers_in_existing_neutron_network_no_floating_ips.yaml --parameters="image=base_image;flavor=m1.small;net_id=da3b99ed-a806-4a9e-9a3d-1aa7ebdc0cd5;subnet_id=010f0972-e28b-44fd-b759-5b2288022111;key_name=key_b-compute04" | 08:28 |
| cmyster | chen: continuing from therve's question and which repo ? Havana official | 08:28 |
| chen | therve no.... | 08:28 |
| therve | chen, Then we can't help you | 08:29 |
| shardy | chen: I assume it's RDO? rpm -qa | grep heat | fpaste | 08:29 |
| cmyster | might be skraynev, I don't remember when it go in | 08:29 |
| skraynev | cmyster: it was merged 11 Nov 2013 (according to git history) | 08:29 |
| skraynev | cmyster: and i have not seen backport to havana | 08:30 |
| *** pas-ha has joined #heat | 08:30 | |
| cmyster | well there you go :) | 08:30 |
| *** duncanjw has joined #heat | 08:30 | |
| chen | shardy therve cmyster : http://paste.openstack.org/show/74320/ | 08:30 |
| chen | skraynev o.....so, what template file should I use ? | 08:31 |
| cmyster | chen: I guess as skraynev its too old and does not have it supported. | 08:31 |
| cmyster | as skraynev said* | 08:32 |
| cmyster | teehee skraynev is probably not THAT old | 08:32 |
| chen | cmyster e... so, I have to install from source code ? | 08:32 |
| skraynev | cmyster: lol =) | 08:32 |
| skraynev | chen: I see two solution: | 08:32 |
| cmyster | chen: Icehouse is not out yet, but you might give it a go via devstack | 08:33 |
| shardy | chen: there are RDO preview packages for icehouse | 08:33 |
| cmyster | depends on what you want to do. for a short POC it might be enough | 08:33 |
| skraynev | chen: first is use master of devstack (heat) | 08:33 |
| cmyster | ya or that | 08:33 |
| shardy | http://repos.fedorapeople.org/repos/openstack/openstack-icehouse/ | 08:33 |
| chen | o... I just want to have a try, to understand what is heat used for at this time.... | 08:34 |
| skraynev | ou.. shardy says truth for RDO | 08:34 |
| cmyster | then either should be fine imo. RDO might be quicker | 08:34 |
| chen | ok. let me have a try. | 08:35 |
| cmyster | and now for something completely different. this template is not producing any output... ideas? http://paste.openstack.org/show/74321/ | 08:35 |
| skraynev | chen: from other side you could try "old" templates | 08:35 |
| chen | skraynev o, would like to try the "old" | 08:36 |
| skraynev | which were added before havana release. | 08:36 |
| chen | skraynev so do need to change repos, no affection to other services | 08:37 |
| skraynev | unfortunately you should check a date of each template in this case | 08:37 |
| *** jistr has joined #heat | 08:37 | |
| cmyster | chen: you can have a look at https://github.com/openstack/heat-templates/blob/master/cfn/F17 there are pretty 'old' templates there | 08:37 |
| skraynev | agree with cmyster about relevant status these templates | 08:38 |
| cmyster | chen: https://github.com/openstack/heat-templates/blob/master/cfn/F17/getting_started.template might be perfect | 08:39 |
| skraynev | chen: change repo of what? if only heat... it may be dangerous, because heat use other clients which should be updated in this case too | 08:39 |
| *** alexheneveld has joined #heat | 08:40 | |
| chen | cmyster It is a template of AWS..... :( | 08:40 |
| shardy | chen: For HOT I'd suggest https://github.com/openstack/heat-templates/blob/master/hot/F20/WordPress_Native.yaml | 08:40 |
| shardy | Try that with the official Fedora 20 cloud image: | 08:40 |
| cmyster | chen: well you did say: chen > o... I just want to have a try, to understand what is heat used for at this time.... | 08:41 |
| shardy | http://openstack.redhat.com/forum/discussion/959/get-fedora-20-now/p1 | 08:41 |
| chen | cmyster y... | 08:41 |
| cmyster | ya | 08:41 |
| *** ifarkas has joined #heat | 08:42 | |
| skraynev | shardy: related question about F20. AFAIK heat is included in devstack by default. But we still should set desired images or download them in glance manually, right? | 08:42 |
| *** chandankumar_ has joined #heat | 08:42 | |
| chen | Thanks guys. I would try WordPress_Native.yaml ! | 08:43 |
| shardy | skraynev: http://docs.openstack.org/developer/heat/getting_started/on_devstack.html | 08:43 |
| shardy | You can just update IMAGE_URLS and devstack will install the F20 image for you | 08:43 |
| shardy | or download them and add to glance if you prefer :) | 08:43 |
| skraynev | shardy: gotcha, just dream about default F20 instead centose images in glance without additional changes in localrc file ;) | 08:44 |
| *** chen has quit IRC | 08:44 | |
| shardy | skraynev: just export it all from your login profile instead of localrc and your dream will come true ;) | 08:51 |
| skraynev | shardy: thumbs up! :) | 08:52 |
| *** liang has joined #heat | 08:53 | |
| *** TonyBurn has joined #heat | 08:55 | |
| *** duncanjw has quit IRC | 08:56 | |
| *** tspatzier has quit IRC | 08:57 | |
| *** tomek_adamczewsk has quit IRC | 09:00 | |
| *** duncanjw has joined #heat | 09:01 | |
| *** tomek_adamczewsk has joined #heat | 09:07 | |
| cmyster | guys, regarding my earlier question, is the template legit ? | 09:08 |
| *** derekh has joined #heat | 09:11 | |
| *** tspatzier has joined #heat | 09:12 | |
| *** gokrokve has joined #heat | 09:13 | |
| *** gokrokve has quit IRC | 09:13 | |
| *** gokrokve has joined #heat | 09:14 | |
| openstackgerrit | Jun Jie Nan proposed a change to openstack/heat: Adjust user_data_format for server with deployment https://review.openstack.org/82992 | 09:14 |
| *** tspatzier has quit IRC | 09:15 | |
| *** gokrokve has quit IRC | 09:16 | |
| *** gokrokve has joined #heat | 09:17 | |
| *** liang has quit IRC | 09:17 | |
| *** liang has joined #heat | 09:17 | |
| *** gokrokve has quit IRC | 09:21 | |
| shardy | cmyster: template looks OK to me, investigating as I also see no outputs | 09:22 |
| cmyster | smells buggish | 09:23 |
| *** Trozz has quit IRC | 09:28 | |
| *** Trozz has joined #heat | 09:29 | |
| *** Trozz_ has joined #heat | 09:30 | |
| *** Trozz_ has quit IRC | 09:30 | |
| *** Trozz has quit IRC | 09:30 | |
| openstackgerrit | Thomas Herve proposed a change to openstack/heat: Raise and catch a specific error during validation https://review.openstack.org/71287 | 09:31 |
| *** Trozz has joined #heat | 09:31 | |
| *** Trozz_ has joined #heat | 09:31 | |
| *** mich5225 has joined #heat | 09:32 | |
| *** Trozz has quit IRC | 09:33 | |
| *** Trozz_ has quit IRC | 09:33 | |
| *** mich5225 has quit IRC | 09:33 | |
| *** Trozz has joined #heat | 09:33 | |
| *** giulivo has joined #heat | 09:39 | |
| *** e0ne has joined #heat | 09:39 | |
| *** Trozz has quit IRC | 09:39 | |
| *** Trozz has joined #heat | 09:40 | |
| *** Trozz has quit IRC | 09:40 | |
| *** qinjin has quit IRC | 09:41 | |
| *** derekh has quit IRC | 09:42 | |
| *** derekh has joined #heat | 09:42 | |
| *** Trozz has joined #heat | 09:43 | |
| *** gokrokve has joined #heat | 09:44 | |
| *** Trozz has quit IRC | 09:45 | |
| openstackgerrit | Thomas Herve proposed a change to openstack/python-heatclient: Skip timeout_mins when not specify https://review.openstack.org/83017 | 09:48 |
| *** tomek_adamczewsk has quit IRC | 09:54 | |
| *** tomek_adamczewsk has joined #heat | 09:55 | |
| *** alexheneveld_ has joined #heat | 10:00 | |
| *** alexheneveld has quit IRC | 10:00 | |
| *** alexheneveld_ is now known as alexheneveld | 10:00 | |
| *** faramir1 has joined #heat | 10:06 | |
| *** nanjj has quit IRC | 10:07 | |
| *** matsuhashi has quit IRC | 10:09 | |
| *** tspatzier has joined #heat | 10:11 | |
| *** tspatzier has quit IRC | 10:15 | |
| *** DaveJ__ has joined #heat | 10:16 | |
| *** chandankumar_ has quit IRC | 10:22 | |
| *** alexheneveld has quit IRC | 10:25 | |
| *** mkollaro has joined #heat | 10:25 | |
| *** Trozz has joined #heat | 10:31 | |
| *** Trozz has quit IRC | 10:34 | |
| *** matsuhashi has joined #heat | 10:39 | |
| *** Trozz has joined #heat | 10:41 | |
| *** Trozz has quit IRC | 10:42 | |
| *** networkn8 has joined #heat | 10:44 | |
| *** Trozz has joined #heat | 10:44 | |
| *** Trozz has quit IRC | 10:45 | |
| *** tomek_adamczewsk has quit IRC | 10:45 | |
| *** mojorison has quit IRC | 10:50 | |
| *** alexheneveld_ has joined #heat | 10:53 | |
| *** che-arne has joined #heat | 10:54 | |
| *** mkollaro has quit IRC | 10:54 | |
| *** tomek_adamczewsk has joined #heat | 10:54 | |
| *** dmakogon_ is now known as denis_makogon | 11:03 | |
| *** julienvey has joined #heat | 11:07 | |
| *** liang has quit IRC | 11:08 | |
| *** julienvey has quit IRC | 11:08 | |
| *** duncanjw has quit IRC | 11:12 | |
| *** rpothier_ has quit IRC | 11:13 | |
| *** saju_m has joined #heat | 11:21 | |
| *** fandi has quit IRC | 11:21 | |
| *** e0ne has quit IRC | 11:23 | |
| *** Trozz has joined #heat | 11:23 | |
| *** Trozz has quit IRC | 11:23 | |
| *** e0ne has joined #heat | 11:23 | |
| *** tspatzier has joined #heat | 11:25 | |
| *** tspatzier has quit IRC | 11:26 | |
| *** e0ne has quit IRC | 11:29 | |
| *** Trozz has joined #heat | 11:29 | |
| *** Trozz has quit IRC | 11:36 | |
| *** ppetit has joined #heat | 11:38 | |
| *** julienvey has joined #heat | 11:41 | |
| *** gokrokve has quit IRC | 11:45 | |
| *** gokrokve has joined #heat | 11:46 | |
| *** nkhare has quit IRC | 11:47 | |
| *** gokrokve has quit IRC | 11:50 | |
| openstackgerrit | Sergey Kraynev proposed a change to openstack/heat: Adding validation algorithm for get attr functions https://review.openstack.org/82488 | 11:52 |
| *** gokrokve has joined #heat | 11:57 | |
| *** jstrachan has quit IRC | 11:59 | |
| *** e0ne has joined #heat | 12:00 | |
| *** Trozz has joined #heat | 12:02 | |
| *** Trozz has quit IRC | 12:03 | |
| *** julienvey has quit IRC | 12:03 | |
| *** wchrisj has joined #heat | 12:04 | |
| *** e0ne has quit IRC | 12:04 | |
| *** ppetit has quit IRC | 12:04 | |
| *** e0ne has joined #heat | 12:05 | |
| *** Slower has quit IRC | 12:05 | |
| *** sgordon has joined #heat | 12:08 | |
| *** fandi has joined #heat | 12:09 | |
| *** duncanjw has joined #heat | 12:09 | |
| *** e0ne has quit IRC | 12:10 | |
| *** matsuhashi has quit IRC | 12:10 | |
| *** matsuhashi has joined #heat | 12:11 | |
| *** julienvey has joined #heat | 12:12 | |
| *** matsuhas_ has joined #heat | 12:13 | |
| *** matsuhashi has quit IRC | 12:13 | |
| *** e0ne has joined #heat | 12:14 | |
| *** aweiteka has joined #heat | 12:14 | |
| *** mkollaro has joined #heat | 12:15 | |
| *** lipinski has joined #heat | 12:18 | |
| *** dims_ has quit IRC | 12:18 | |
| *** julienvey has quit IRC | 12:18 | |
| *** mkollaro has quit IRC | 12:21 | |
| *** dims_ has joined #heat | 12:21 | |
| *** e0ne_ has joined #heat | 12:23 | |
| *** duncanjw has quit IRC | 12:23 | |
| *** e0ne has quit IRC | 12:23 | |
| *** e0ne has joined #heat | 12:25 | |
| *** david-lyle has quit IRC | 12:27 | |
| *** rpothier_ has joined #heat | 12:27 | |
| *** e0ne_ has quit IRC | 12:28 | |
| openstackgerrit | A change was merged to openstack/heat: Fix creating docker containers https://review.openstack.org/82457 | 12:34 |
| *** mkollaro has joined #heat | 12:36 | |
| *** julienvey has joined #heat | 12:37 | |
| *** ppetit has joined #heat | 12:48 | |
| sdake_ | morning | 12:52 |
| *** gokrokve has quit IRC | 12:53 | |
| chmouel | sdake_: good morning | 12:53 |
| *** gokrokve has joined #heat | 12:53 | |
| *** gokrokve has quit IRC | 12:57 | |
| therve | chmouel, I asked a question on https://bugs.launchpad.net/heat/+bug/1297122 btw | 12:57 |
| uvirtbot | Launchpad bug 1297122 in heat "Docker plugin name conflict with docker library" [Undecided,In progress] | 12:57 |
| *** tomek_adamczewsk has quit IRC | 12:59 | |
| chmouel | therve: cool ansered there (i get too much bug reports floating by my mailbox to gets the one directly answered to my requests) | 13:02 |
| *** tomek_adamczewsk has joined #heat | 13:02 | |
| therve | Hum okay | 13:02 |
| therve | chmouel, __import__ then | 13:02 |
| chmouel | isn't that a bit hacky? | 13:03 |
| therve | Renaming is disruptive. Python supports namespaces just fine :) | 13:03 |
| therve | You can use importlib too | 13:03 |
| chmouel | by just fine you mean just fine in python3 ? | 13:05 |
| chmouel | can do the importlib I thought that module wasn't working at first so i didn't expect much people using it | 13:06 |
| chmouel | and the renaming would be harmless | 13:06 |
| *** e0ne_ has joined #heat | 13:07 | |
| *** gokrokve has joined #heat | 13:07 | |
| *** matsuhas_ has quit IRC | 13:07 | |
| therve | No I mean just fine in general | 13:08 |
| therve | chmouel, You break documentation too | 13:10 |
| *** e0ne has quit IRC | 13:10 | |
| chmouel | yeah saw that, well i guess i'd use the 'fine' importlib :) | 13:10 |
| therve | I wonder if that works in that case though | 13:11 |
| therve | Maybe you do need renaming? We don't install it in the "docker" namespace? | 13:11 |
| chmouel | I don't think so, it only get confused during the testrunning | 13:12 |
| chmouel | may need some invstigation | 13:12 |
| *** achampion has quit IRC | 13:12 | |
| *** arbylee has joined #heat | 13:14 | |
| *** alexheneveld_ has quit IRC | 13:14 | |
| *** blomquisg has joined #heat | 13:14 | |
| *** arbylee has quit IRC | 13:16 | |
| *** arbylee has joined #heat | 13:18 | |
| *** arbylee has quit IRC | 13:18 | |
| *** arbylee has joined #heat | 13:19 | |
| *** tspatzier has joined #heat | 13:19 | |
| *** faramir1 has quit IRC | 13:20 | |
| *** hazmat has quit IRC | 13:21 | |
| *** rbuilta has joined #heat | 13:22 | |
| *** tspatzier has quit IRC | 13:28 | |
| *** jcru has joined #heat | 13:29 | |
| *** arbylee has quit IRC | 13:30 | |
| *** arbylee has joined #heat | 13:30 | |
| *** nosnos has quit IRC | 13:35 | |
| sdake_ | anyone know how I would go about running the databsae migrations | 13:40 |
| zaneb | sdake_: heat-manage --help | 13:41 |
| sdake_ | zaneb thanks :) | 13:41 |
| chmouel | :) | 13:41 |
| *** samstav has joined #heat | 13:41 | |
| * IgorYozhikov is now away: went away... | 13:43 | |
| *** IgorYozhikov is now known as IYozhikov_away | 13:43 | |
| *** lazy_prince is now known as killer_prince | 13:46 | |
| *** asalkeld has joined #heat | 13:48 | |
| skraynev | guys, should default value (which defined in property schema) be used during update? | 13:51 |
| skraynev | f.e. I have property with default = [] | 13:51 |
| *** ppetit_ has joined #heat | 13:51 | |
| *** ppetit_ has quit IRC | 13:51 | |
| skraynev | 1. create stack with [val1, val2] | 13:51 |
| skraynev | 2. update stack and don't mention this value in template | 13:52 |
| skraynev | in this case prop_diff, thinks that value is None, instead [] | 13:52 |
| skraynev | zaneb, shardy ^^ | 13:53 |
| *** randallburt has joined #heat | 13:53 | |
| *** randallburt has joined #heat | 13:53 | |
| zaneb | skraynev: that seems like bug | 13:54 |
| zaneb | *a* bug | 13:54 |
| skraynev | zaneb: hm. so expected behavior is use default value during update, right? | 13:54 |
| *** ppetit has quit IRC | 13:55 | |
| *** vijendar has joined #heat | 13:56 | |
| zaneb | skraynev: that makes sense to me | 13:57 |
| skraynev | thx, will ask stevebaker also ) | 13:58 |
| *** skraynev is now known as skraynev_going_h | 13:59 | |
| *** skraynev_going_h is now known as skraynev_afk | 13:59 | |
| *** achampion has joined #heat | 14:00 | |
| pas-ha | guys, can someone point me to a chat where to ask devstack-related questions? | 14:00 |
| larsks | pas-ha: Possibly #openstack. | 14:00 |
| sdake_ | pas-ha #openstack-dev | 14:01 |
| pas-ha | sdake_, larsks: thnx | 14:01 |
| sdake_ | I don't think tehre is a devstack specific channel | 14:01 |
| sdake_ | pas-ha https://wiki.openstack.org/wiki/IRC for future reference | 14:01 |
| dims_ | stevebaker, (or anyone else). OpenStackClients lookups a nova service from the keystone catalog that auth_url points to. When multi_cloud is switched on, i can't seem to find where the auth_url is set in the context based on the headers of the incoming request | 14:02 |
| *** tspatzier has joined #heat | 14:02 | |
| *** IlyaE has joined #heat | 14:03 | |
| *** vijendar has quit IRC | 14:03 | |
| sdake_ | dims| someone complained about that in a review - but the review was for documentation changes | 14:03 |
| sdake_ | so I'd suggest filing a bug | 14:03 |
| sdake_ | so it gets tracked | 14:03 |
| *** cmyster has quit IRC | 14:03 | |
| *** vijendar has joined #heat | 14:03 | |
| *** samstav has quit IRC | 14:04 | |
| *** andersonvom has joined #heat | 14:04 | |
| shardy | zaneb, sdake: interested to get your feedback on bug #1297761 when you get a moment | 14:04 |
| uvirtbot | Launchpad bug 1297761 in heat "No validation of top-level template keys" [Undecided,New] https://launchpad.net/bugs/1297761 | 14:05 |
| shardy | follows on from the resource-level custom-keys discussion, turns out we don't validate the top-level keys at all | 14:05 |
| shardy | I thought we did, or at least at one point we did.. | 14:05 |
| dims_ | sdake, will do, also another problem is get_admin_context has no clue about auth_url as well it always picks up from config so periodic watcher has no clue either | 14:05 |
| sdake_ | dims_ tbh I dont' quite understand what is required to fix it, but I know the review in question where it was brought up was only doc changes, so it was +2'ed | 14:06 |
| *** varora has left #heat | 14:06 | |
| therve | shardy, Did we? | 14:06 |
| sdake_ | so getting it in the bug tracker would help people discuss it, unless you have a solution ready to rock :) | 14:06 |
| therve | Validation has always been quite loose :) | 14:06 |
| dims_ | sdake, yep! | 14:07 |
| zaneb | shardy: yeah, we probably didn't ever do that in HOT | 14:07 |
| zaneb | shardy: I thought we did for cfn templates though? | 14:07 |
| shardy | zaneb: I've tested and neither format raises an error | 14:08 |
| shardy | at least on Icehouse, I've only tested HOT on Havana | 14:08 |
| zaneb | it's probably the same then | 14:08 |
| shardy | I may start a ML thread about it, as it's pretty unhelpful to user to silently accept malformed templates, but I'm worried folks may be making use of the current behavior | 14:09 |
| * shardy is thinking of the template-metadata stuff which was previously discussed and rejected | 14:09 | |
| shardy | turns out we allowed it all along :\ | 14:09 |
| *** duncanjw has joined #heat | 14:09 | |
| zaneb | shardy: I don't think anybody will be doing superset-stuff, because the API of the Template class doesn't allow you to access it | 14:10 |
| zaneb | shardy: shhhhhhh | 14:10 |
| sdake_ | i think hot is pretty well described in the spec | 14:10 |
| shardy | zaneb: Yeah, that is what I'm hoping | 14:10 |
| sdake_ | people operating outside of the spec - well atleast they have a spec they SHOULD be using | 14:10 |
| sdake_ | if they are using things that are bugs, essentially exploiting heat, I really don't have a problem with their templates breaking | 14:10 |
| shardy | Ok I'll just post a patch fixing it and we can discuss in the review | 14:11 |
| zaneb | sdake_: agree but we _will_ need a db migration to clean up their templates | 14:11 |
| sdake_ | our parser should be very pendantic imo :) | 14:11 |
| zaneb | +1 | 14:11 |
| sdake_ | zaneb yes, I'm working on that ? | 14:11 |
| sdake_ | as we discussed yesterday | 14:11 |
| shardy | sdake_: well I kinda agree, but e.g tspatzier's recent patch for custom resource keys shows there is already some (ab)use of our lax validation | 14:12 |
| sdake_ | I guess I dont know all the conditions | 14:12 |
| shardy | tspatzier: Ok, you're here ^^ | 14:12 |
| shardy | any feedback appreciated before I break everybody and get shouted at (again) ;) | 14:12 |
| sdake_ | shardy yes and I think we want to draw a clear line in the sand about that :) | 14:12 |
| zaneb | sdake_: maybe add deleting extraneous keys to that migration? | 14:12 |
| tspatzier | hey shardy, what is the discussion? | 14:12 |
| sdake_ | zaneb I was going to make some separate migrations | 14:12 |
| tspatzier | whether we should be really strict? | 14:12 |
| sdake_ | since they are separeate bugs | 14:12 |
| *** samstav has joined #heat | 14:12 | |
| shardy | tspatzier: It's about strict enforcement/validation of top-level template keys (sections) | 14:13 |
| shardy | tspatzier: e.g atm we silently allow a mis-spelt resources or outputs section | 14:13 |
| shardy | ref https://bugs.launchpad.net/heat/+bug/1297761 | 14:13 |
| uvirtbot | Launchpad bug 1297761 in heat "No validation of top-level template keys" [Undecided,New] | 14:13 |
| sdake_ | strict enforcement is necessary, the fact that we aren't strict now is because the parsers need love, not because that is our intention :) | 14:13 |
| shardy | tspatzier: my opinion is we should do strict validation and reject malformed section keys, but given your recent custom-keys patch I'd like to discuss it first :) | 14:14 |
| tspatzier | abandoned my patch already. I am all for strict validation | 14:14 |
| shardy | tspatzier: Ok, I'll post a patch which enforces section validation, thanks | 14:15 |
| *** cfriesen has joined #heat | 14:15 | |
| tspatzier | shardy: we have discussed internally and I am fine with strict checking. Let's do it now to not cause further confusion going forward. | 14:15 |
| shardy | tspatzier: +1 | 14:15 |
| sdake_ | +1 lets have pain soone rrather then later | 14:15 |
| tspatzier | if something else is needed, let's add it to HOT in a coordinated fashion, i.e. BP, discuss, document ... | 14:15 |
| sdake_ | hot has really turned out nicely | 14:16 |
| sdake_ | I was suspect early on :) | 14:16 |
| *** Darnoth has joined #heat | 14:16 | |
| sdake_ | nice job tspatzier :) | 14:16 |
| tspatzier | sdake_, yep, I also like it taking off | 14:16 |
| *** alexheneveld has joined #heat | 14:17 | |
| tspatzier | thanks sdake_ | 14:17 |
| tspatzier | shardy: so are you working on that bug you linked above? | 14:18 |
| shardy | tspatzier: I was about to start, unless you'd like to take it :) | 14:20 |
| shardy | looked at some code but not really written a fix or tests yet | 14:20 |
| shardy | (it was only reported this morning) | 14:20 |
| dims_ | sdake, fyi - https://bugs.launchpad.net/heat/+bug/1297880 | 14:21 |
| uvirtbot | Launchpad bug 1297880 in heat "multi_cloud : looking up services from keystone catalog - possible issues" [Undecided,New] | 14:21 |
| *** sabeen has joined #heat | 14:22 | |
| tspatzier | shardy: I could look into it. Have to see what the best code is, but the logic would be: everything that is in hot_spec.rst will pass, everything else will be rejected. | 14:22 |
| shardy | I think it's fairly simple, we just need to raise an exception if *Template is created with a key not in SECTIONS as defined in each class | 14:23 |
| tspatzier | shardy: right, that's what I thought | 14:23 |
| shardy | and work out a place we can do it (e.g on create) which won't break every existing template stored in the DB if it contains invalid keys | 14:24 |
| shardy | or have a migration which removes all invalid keys I guess | 14:24 |
| tspatzier | hm, have to think about the existing templates in the DB | 14:24 |
| sdake_ | shardy migrations are what we are doing imo :) | 14:25 |
| tspatzier | yes, we did it for the last syntax checking patches | 14:25 |
| sdake_ | we need a migration for version and parameters too | 14:25 |
| sdake_ | if I can figure out how to get an older version of heat running, I'd submit some patches :) | 14:26 |
| tspatzier | shardy: so I take this bug if you like. I guess you like, right ;-) | 14:29 |
| randallburt | sdake_: going to run a spell checker in a sqlalchemy migration? ;) | 14:29 |
| chmouel | that should not just crash right http://pastie.org/pastes/8970418/text?key=w13retcbbb3vsh99t2cya it should come back something to the user? | 14:29 |
| *** rcleere has joined #heat | 14:30 | |
| *** tomek_adamczewsk has quit IRC | 14:30 | |
| randallburt | tspatzier: we've been seeing failed calls to stack-list when we show deleted stacks because of the lack of validation earlier on. Couldn't we just not validate on stack-list/stack-show? | 14:31 |
| sdake_ | randallburt I investigated that | 14:31 |
| sdake_ | it is untidy at best | 14:31 |
| tspatzier | randallburt: so you mean if we do strict checking on create, we never have to validate again? | 14:32 |
| randallburt | tspatzier: and UPDATE too, but sure. | 14:32 |
| sdake_ | the code base doesn't support that model easily | 14:32 |
| *** tomek_adamczewsk has joined #heat | 14:33 | |
| tspatzier | so for HOT, most is done in the translate code today. Parameters are done in the param schema code. | 14:33 |
| randallburt | sdake_: Yeah, I imagine instantiating the full business object for lists and shows is a bit problemmatic, however, we already re-validate before every operation, so I wonder if its just not removing the calls to validate in constructors and whatnot. | 14:33 |
| sdake_ | whenever a stack object is instantiated, (on any API operation) the entire thing is parsed | 14:33 |
| *** vijendar has quit IRC | 14:33 | |
| sdake_ | it is not removing validate calls | 14:33 |
| sdake_ | there is no validate call to remotge | 14:33 |
| sdake_ | to remove | 14:33 |
| sdake_ | when you do something like stack list, the template is parsed | 14:34 |
| shardy | tspatzier: If you have time to work on it, feel free :) | 14:34 |
| sdake_ | the template parsing doesn't call "hey validate this" it does validation inline with parsing | 14:34 |
| randallburt | sdake_ template_format.parse still validates keys? | 14:34 |
| randallburt | sdake_: poop. | 14:34 |
| tspatzier | shardy: ok, I'll take it | 14:34 |
| shardy | tspatzier: thanks! | 14:34 |
| sdake_ | perhaps a better long term solution is to provide a separate validate operation independent of the parsing | 14:35 |
| sdake_ | but I think given the number of days we have left to fix that, that would be pretty invasive | 14:35 |
| sdake_ | eg, two steps -> validate, then parse | 14:35 |
| sdake_ | the parser always assumes the validation is correct | 14:35 |
| sdake_ | right now parsing and validation are jammed together | 14:35 |
| *** e0ne has joined #heat | 14:36 | |
| sdake_ | randallburt I am just not keen to essentially rewrite the parsers to support that model for rc1 | 14:37 |
| randallburt | sdake_: yeah, def sounds tricky. just a thought, though. For now, data migrations catch *most* things, but when stuff is misspelled or in the wrong place, it can be hard to detect/fix. | 14:37 |
| jasond | sdake_: i noticed that. made fixing difficult https://bugs.launchpad.net/heat/+bug/1296747 ended up having to remove validation for nova keypair | 14:37 |
| uvirtbot | Launchpad bug 1296747 in heat "Nova keypair cannot be updated" [High,Fix committed] | 14:37 |
| randallburt | sdake_: agreed, | 14:37 |
| tspatzier | sdake_: agree on the long term cleanup. For now, it would do a minimal fix for the bug I guess. | 14:37 |
| sdake_ | tspatzier which bug are you talking about | 14:37 |
| sdake_ | so far, there are two I knwo of | 14:37 |
| sdake_ | people can specify wacky version #s, which breaks delete/list | 14:38 |
| tspatzier | The one I just inherited from shardy: https://bugs.launchpad.net/heat/+bug/1297761 | 14:38 |
| uvirtbot | Launchpad bug 1297761 in heat "No validation of top-level template keys" [Undecided,New] | 14:38 |
| sdake_ | people can specify mixed hot/cfn parameters which breaks delete/list | 14:38 |
| *** alexpilotti has joined #heat | 14:39 | |
| tspatzier | sdake_: the parameter syntax should have been fixed a few weeks back by some schema code refactoring we did. | 14:39 |
| *** daveed has quit IRC | 14:39 | |
| *** e0ne_ has quit IRC | 14:39 | |
| sdake_ | tspatzier, if people have old templates in their db with mixed cfn/hot parameters, they cannot be listed/deleted | 14:39 |
| tspatzier | sdake_: understand. But there have been migrate db patches coming in. | 14:40 |
| sdake_ | tspatzier there is no migrate patch for params | 14:40 |
| sdake_ | working on that atm | 14:40 |
| tspatzier | I thought andrew_plunk was working on one. | 14:41 |
| sdake_ | tspatzier I am working on https://bugs.launchpad.net/bugs/1296859 - actually | 14:42 |
| uvirtbot | Launchpad bug 1296859 in heat "stack-list throws an error if any stack returned contains a template error" [Medium,Triaged] | 14:42 |
| *** chandan_kumar has quit IRC | 14:42 | |
| sdake_ | tspatzier I don't see any params bug in rc1 assigned to andrew_plunk | 14:42 |
| sdake_ | but I think we can't backport db migrations, so ideally this is something we need to fix before rc1 | 14:42 |
| tspatzier | This is the migrate script I remember: https://github.com/openstack/heat/blob/master/heat/db/sqlalchemy/migrate_repo/versions/037_migrate_hot_template.py | 14:42 |
| sdake_ | cool so looks like parameters is handled | 14:43 |
| tspatzier | sdake_: your bug is re template version. For that one I cannot remember of a db script. | 14:44 |
| *** skraynev_afk is now known as skraynev | 14:44 | |
| sdake_ | tspatzier right | 14:44 |
| sdake_ | there is none definately for version | 14:44 |
| tspatzier | resources and outputs should be handled as well: the 041 script | 14:44 |
| sdake_ | cool so we are all set then once I finish up | 14:44 |
| *** chandankumar_ has joined #heat | 14:45 | |
| tspatzier | right, and I will take care of the top level keys in course of fixed the bug I took | 14:45 |
| tspatzier | s/fixed/fixing/ | 14:45 |
| *** chandankumar_ has quit IRC | 14:46 | |
| *** mestery has joined #heat | 14:46 | |
| *** chandan_kumar has joined #heat | 14:47 | |
| *** jmckind has joined #heat | 14:48 | |
| *** chandan_kumar has quit IRC | 14:48 | |
| *** chandan_kumar has joined #heat | 14:48 | |
| tspatzier | sdake_, shardy: if I throw an error message in the fix, is there some i18n lock-down I should be aware of? | 14:49 |
| sdake_ | we are in string freeze | 14:50 |
| tspatzier | so no new error message? just throw e.g. a KeyError? | 14:51 |
| shardy | Hmm, I was thinking we'd want a new InvalidTemplateSection exception | 14:52 |
| tspatzier | shardy: would that involve translation? The exception name itself is very self-explaining | 14:53 |
| *** andersonvom_ has joined #heat | 14:53 | |
| *** david-lyle has joined #heat | 14:53 | |
| *** radez_g0n3 is now known as radez | 14:54 | |
| shardy | tspatzier: well this is the patch I'd just started, but it would involve a new string | 14:54 |
| shardy | http://paste.fedoraproject.org/88795/45593139/ | 14:54 |
| shardy | I wonder if we can just ask for a string-freeze exception for it | 14:54 |
| shardy | better to raise a descriptive error in one language than a non-descriptive one in all of them imo | 14:54 |
| *** chandan_kumar has quit IRC | 14:55 | |
| tspatzier | shardy: agree | 14:55 |
| shardy | If that's not allowed then I guess raise a generic error and we fix in Juno | 14:55 |
| *** andersonvom has quit IRC | 14:55 | |
| tspatzier | shardy: I can include this in my fix. If people reject, we do the generic error thing. | 14:56 |
| shardy | +1 | 14:56 |
| *** chandan_kumar has joined #heat | 14:57 | |
| *** bvandenh has quit IRC | 15:02 | |
| *** chandan_kumar has quit IRC | 15:03 | |
| *** jstrachan has joined #heat | 15:04 | |
| *** mestery has quit IRC | 15:05 | |
| *** tspatzier has quit IRC | 15:07 | |
| openstackgerrit | Anderson Mesquita proposed a change to openstack/heat: Add OS:Barbican:Order resource https://review.openstack.org/81906 | 15:08 |
| openstackgerrit | Anderson Mesquita proposed a change to openstack/heat: Add OS:Barbican:Secret resource https://review.openstack.org/79355 | 15:08 |
| *** chandan_kumar has joined #heat | 15:09 | |
| *** andrew_plunk has joined #heat | 15:11 | |
| *** jstrachan has quit IRC | 15:13 | |
| *** andersonvom_ has quit IRC | 15:14 | |
| *** andersonvom has joined #heat | 15:14 | |
| *** ppetit has joined #heat | 15:16 | |
| *** wchrisj has quit IRC | 15:18 | |
| *** lsmola_ has quit IRC | 15:19 | |
| *** pas-ha has quit IRC | 15:19 | |
| *** tspatzier has joined #heat | 15:23 | |
| *** samstav has quit IRC | 15:25 | |
| openstackgerrit | Zane Bitter proposed a change to openstack/heat: Fail if non-existent security group referenced https://review.openstack.org/81558 | 15:27 |
| *** lsmola_ has joined #heat | 15:34 | |
| *** randallburt has quit IRC | 15:34 | |
| *** Linz has joined #heat | 15:35 | |
| *** ifarkas has quit IRC | 15:36 | |
| lipinski | Is there a way to convert a string to a json object or separated list? | 15:36 |
| lipinski | I'm trying to pass a list (which is an output of one resource) to the networks property of a server ({ network: 'public' }), but the engine is recognizing it as a string, not a Map | 15:37 |
| *** Linz has quit IRC | 15:39 | |
| *** ifarkas has joined #heat | 15:40 | |
| *** saju_m has quit IRC | 15:43 | |
| *** Linz has joined #heat | 15:44 | |
| shardy | lipinski: You can specify an index to select the item you want from the list via get_attr: | 15:45 |
| shardy | http://docs.openstack.org/developer/heat/template_guide/hot_spec.html#intrinsic-functions | 15:45 |
| lipinski | shardy: I have a resource with an output of obj: { network: "public" } | 15:45 |
| shardy | lipinski: which resource? | 15:46 |
| *** spzala has joined #heat | 15:46 | |
| lipinski | I try to pass that to a networks property of a OS::Nova::Server via networks:[ { get_attr: [ my_res, obj ] } ] | 15:46 |
| lipinski | shardy: a provider resource | 15:46 |
| lipinski | I get: networks Property error : networks: 0 "{u'network': u'public'}" is not a map | 15:46 |
| lipinski | So, it is getting the data I want. It's just that it seems to be regocnizing the result ot get_attr as a string and not interpreting it as a JSON | 15:47 |
| *** ifarkas has quit IRC | 15:48 | |
| therve | lipinski, Can you work around by just returning "public"? | 15:48 |
| therve | I think attributes have to be strings for now | 15:48 |
| *** samstav has joined #heat | 15:48 | |
| lipinski | therve: it's a deep problem that has another issue I'm trying to work around - not being able to determine list sizes nor being able to merge lists | 15:49 |
| lipinski | I'm trying to have one resource build a list of networks that I pass to OS::Nova::Server. The "builder" resource will pass back an arbitrary number of networks | 15:50 |
| *** samstav has quit IRC | 15:50 | |
| lipinski | I was trying to have it return a list, and just pass that onto OS::Nova::Server - but I keep running into this string vs list vs map/json issue. | 15:50 |
| lipinski | I was hoping that since other OS resources can return attributes other than strings that I could too. | 15:51 |
| *** kgriffs_afk is now known as kgriffs | 15:51 | |
| shardy | Looks like all nested stack attributes do get converted to strings: | 15:51 |
| shardy | https://github.com/openstack/heat/blob/master/heat/engine/stack_resource.py#L325 | 15:51 |
| shardy | lipinski: if you can raise a bug with some minimal examples demonstrating what you're trying to do that would be most helpful | 15:52 |
| *** wchrisj has joined #heat | 15:52 | |
| lipinski | I'm guessing that provider resources are then subclasses of stack resources | 15:52 |
| shardy | lipinski: yup, provider resources are nested stacks | 15:53 |
| lipinski | When I create a stack from the provide resource itself, the attribute seems to be recognized in the output as a json: | 15:55 |
| lipinski | ... "output_value": { "network": "private" }, ... | 15:55 |
| *** duncanjw has quit IRC | 15:58 | |
| *** killer_prince has quit IRC | 15:59 | |
| *** killer_prince has joined #heat | 15:59 | |
| *** radez is now known as radez_g0n3 | 15:59 | |
| *** mestery has joined #heat | 16:01 | |
| openstackgerrit | Steven Hardy proposed a change to openstack/heat: Don't log exception for None timeout_mins https://review.openstack.org/83109 | 16:02 |
| *** radez_g0n3 is now known as radez | 16:02 | |
| jasond | anybody else want to weigh in on a HOT "mappings" section? https://review.openstack.org/#/c/81918/ | 16:03 |
| shardy | jasond: Is there a reason why you can't use parameter declarations in the environment instead? | 16:05 |
| *** TonyBurn has quit IRC | 16:05 | |
| *** wchrisj has quit IRC | 16:05 | |
| *** lsmola_ has quit IRC | 16:06 | |
| jasond | shardy: the mapping needs to be coupled to the template | 16:06 |
| *** ppetit has quit IRC | 16:06 | |
| shardy | jasond: well an environment is coupled with the template at the API level | 16:07 |
| *** tomek_adamczewsk has quit IRC | 16:07 | |
| *** wchrisj has joined #heat | 16:07 | |
| shardy | jasond: perhaps providing some example use-cases would be helpful? | 16:07 |
| jasond | shardy: did you read through the comments? | 16:07 |
| therve | Yeah I'm not sure I understand the "image name-to-image ID mapping" example | 16:07 |
| therve | image have names already? | 16:07 |
| jasond | therve: http://dunsmor.com/pastebin/1395850105.txt | 16:08 |
| *** tomek_adamczewsk has joined #heat | 16:09 | |
| jasond | therve: the global image name-to-id mappings change | 16:09 |
| *** sgordon has quit IRC | 16:09 | |
| therve | So you hide which images that you use. Hum... | 16:09 |
| jasond | therve: and that greatly reduces the reliability of any user-data scripts | 16:09 |
| jasond | therve: what do you mean? | 16:09 |
| therve | jasond, Why? | 16:09 |
| jasond | therve: because the environment that the userdata script was developed for has changed | 16:10 |
| therve | I mean you don't pass a real image, you pass a name that the template translates. With hardcoded values in your template that make it hard to port | 16:10 |
| shardy | therve: +1 | 16:10 |
| jasond | therve: how does it translate the name? | 16:11 |
| therve | jasond, Sorry I don't understand what you just said | 16:11 |
| shardy | I personally prefer the direction we've been headed, which is to make the stack template as reusable as possible, and have most things which can change defined in the environment | 16:11 |
| *** julienvey has quit IRC | 16:11 | |
| shardy | The smarter constraint stuff makes that even easier, as you can query the service rather than having a hard-coded list | 16:11 |
| *** wchrisj has quit IRC | 16:12 | |
| shardy | As for the name/id thing, I like the approach of making the underlying resource do the lookup by name when we find a name rather than ID | 16:12 |
| shardy | e.g like already happens in some resources | 16:12 |
| jasond | therve: when the teamplate author develops a user-data script, they test it by running it against a particular image. but if the image has changed, that script can easily fail | 16:12 |
| *** randallburt has joined #heat | 16:13 | |
| shardy | jasond: So they define an image ID in their environment, and re-test everything when they make a change to the environment file (which can be reused for all their stack templates) | 16:13 |
| shardy | If you have 1000 stack templates to maintain, surely that's much easier than having to sed the lot with hard-coded id's? | 16:14 |
| therve | jasond, Well okay, so that's solved by having a constraint on the images. Mappings don't change that. | 16:14 |
| shardy | If a mapping definition does get added, I'd much prefer it to be in the environment rather than the template itself | 16:14 |
| jasond | shardy: that's at the heat-level. the mapping needs to be at the template-level | 16:14 |
| *** skraynev is now known as skraynev_afk | 16:14 | |
| shardy | jasond: well my argument is that it doesn't ;) | 16:15 |
| therve | Note that it *may* be useful. I was personally thinking about an "attributes" or "variables" section when you can stuff reusable data | 16:15 |
| radix | good morning | 16:15 |
| shardy | decoupling the declarative model from changes in the environment is exactly what the environment file is for IMO | 16:15 |
| tspatzier | shardy, therve, jasond: just to restate what I wrote in the review: I also think this rather belongs to environment, since in the template it makes the template non-portable. | 16:16 |
| *** julienvey has joined #heat | 16:16 | |
| *** killer_prince has quit IRC | 16:17 | |
| *** killer_prince has joined #heat | 16:18 | |
| *** samstav has joined #heat | 16:19 | |
| *** varora has joined #heat | 16:20 | |
| jasond | shardy, tspatzier, therve: let me try to give an example. if a rackspace customer develops a template with a user-data script, they test it against a specific CentOS 6.5 image. if the Heat environment changes the image id that CentOS 6.5 maps to, their user-data script could break | 16:20 |
| *** lsmola_ has joined #heat | 16:20 | |
| *** e0ne_ has joined #heat | 16:21 | |
| jasond | a key point is that this is a template that the end user maintains/supports | 16:21 |
| *** jstrachan has joined #heat | 16:22 | |
| jasond | they have no control over heat | 16:22 |
| therve | They can still specify an environment though | 16:22 |
| tspatzier | jasond: but if you code the image ID in the template and give the template to somebody else, it won't work if there are different image ID in the other guy's openstack. | 16:22 |
| therve | "if" :) | 16:22 |
| shardy | jasond: can't they maintain their own environment (in addition to whatever you provide via the global environment)? | 16:22 |
| shardy | their environment will always take precedence, so they have control and "the Heat environment" will never change | 16:23 |
| shardy | unless they change it :) | 16:23 |
| tspatzier | agree, shardy. That's how it should be. | 16:23 |
| shardy | as tspatzier says, it's also much better from a reusability perspective | 16:24 |
| *** e0ne has quit IRC | 16:24 | |
| randallburt | why not both? enable an in-template map and have it be overridable via environment? | 16:24 |
| jasond | but then the template wouldn't be complete without the environment file | 16:24 |
| *** duncanjw has joined #heat | 16:25 | |
| *** jstrachan has quit IRC | 16:25 | |
| jasond | it's vital information that would have to be kept in the environment file | 16:25 |
| shardy | imagine they want to use the same template in dev, test and production, uh, environments, with different image id's or paramters controlling network setup, logging etc - having it hard-coded in the template is just a headache | 16:25 |
| tspatzier | jasond: could we do something like: if mapping exists in environment, use it. If not, do lookup by name. Seems like this would need to be done in the resource. | 16:25 |
| shardy | jasond: agree, but that's what the environment is *for* | 16:25 |
| openstackgerrit | Jenkins proposed a change to openstack/heat: Updated from global requirements https://review.openstack.org/76689 | 16:26 |
| jasond | shardy: well it would be an optional feature | 16:27 |
| shardy | IMO it makes no difference to the user if the "vital" information exists in the environment, the template, or on a bit of paper the user types in as parameters, other than defining it in the environement is much more convenient from a reuse perspective | 16:27 |
| andrew_plunk | I like the idea of the environment having the ability to "map" parameter values passed into a template (which can be overridden by a user specified environment). | 16:27 |
| randallburt | tspatzier: seems a little too specific to images and flavors grated thats where the original problems cropped up. | 16:27 |
| shardy | jasond: I guess the point is to promote use of environments to avoid some of the bad points of CFN, such as hard-coded URL's and mappings in their templates | 16:28 |
| randallburt | isn't the level of effort the "same" if the user has to maintain their own environment file or just change a mapping stanza in the template? | 16:28 |
| tspatzier | randallburt: this is re my comment on doing it in the resource? Then maybe not in the general resource class, but in those that need to do those kinds of lookups by name. | 16:28 |
| jasond | randallburt: how would that work? an environment file per template? | 16:29 |
| randallburt | tspatzier: k. just think that whatever solution should be generic enough to satisfy similar issues around values | 16:29 |
| tspatzier | randallburt: sure | 16:29 |
| randallburt | and don't get me wrong; I was never a fan of the mapping section in CFN, but I def understand jasond's concern around the requirement to have a mapping file to make a given template work. | 16:30 |
| randallburt | but moving that mapping from the template to the environment seems same/same to me. | 16:30 |
| randallburt | its still something a user will need to change and maintain as the template itself changes. | 16:31 |
| therve | I'm still not sold on the use case TBH | 16:32 |
| therve | Sure the script can break. But maybe not. | 16:32 |
| *** asalkeld has quit IRC | 16:32 | |
| tspatzier | randallburt: sure, it has to be maintained. But on the other hand, you might also end up editing lots of templates instead of one env as shardy has mentioned before. | 16:32 |
| jasond | therve: we've found out that the script does break, and often | 16:32 |
| tspatzier | I am also not yet convinced we need mappings in the template. This opens the door for ugly mis-uses for hardcoded, environment specific stuff in the template. | 16:33 |
| lipinski | shardy: I opened https://bugs.launchpad.net/heat/+bug/1297938 for the provider attribute string thing. Let me know if this looks sufficient.. | 16:34 |
| uvirtbot | Launchpad bug 1297938 in heat "Provider Resource Attributes forced to string type" [Undecided,New] | 16:34 |
| *** duncanjw has quit IRC | 16:35 | |
| therve | jasond, But do you really solve something by hardcoding the values? Surely it doesn't make it work with the image I want to use | 16:35 |
| shardy | lipinski: lgtm, thanks | 16:35 |
| jasond | therve, tspatzier, shardy: the existence of a user-data script makes the template less portable. the mappings section doesn't make it less portable than it already is | 16:35 |
| lipinski | shardy: thanks. | 16:35 |
| *** wchrisj has joined #heat | 16:36 | |
| tspatzier | jasond: with the new software-orchestration stuff we have a way of living without inlined user-data scripts. | 16:36 |
| shardy | jasond: well that's the problem we've been trying to solve software-config | 16:36 |
| tspatzier | You can actually decouple the software definitions from the infrastructure for different environments. | 16:36 |
| randallburt | tspatzier: that remains to be well proven and is something for the future, not the same problem really | 16:36 |
| therve | randallburt, mappings are a thing for the future :) | 16:37 |
| shardy | jasond: moving in the direction of making the main template declaration more portable, hence my comment that this is a step in the wrong direction, IMHO | 16:37 |
| tspatzier | randallburt: partly agree. But that is the direction. And I would rather make this solid instead of introducing a quick fix. | 16:37 |
| therve | A quick fix that has incidence on the overall template structure | 16:38 |
| therve | You need to cover more use-cases IMHO | 16:38 |
| shardy | jasond: e.g now, you could have a completely reusable top-level declarative model, with all the imperative parts specified via the environment | 16:38 |
| shardy | variables/constants in the environment and e.g abstracted software config via provider resources or pulled in via get_file | 16:39 |
| tspatzier | could this be a topic for the design summit? | 16:39 |
| tspatzier | I mean, the general structure going forward with software config, decoupling from infrastructure etc. | 16:40 |
| tspatzier | just a thought | 16:40 |
| shardy | tspatzier: topic title "quick fixes vs long term roadmap" ;) | 16:40 |
| tspatzier | ;-) | 16:40 |
| shardy | I agree it would be a good topic for discussion | 16:40 |
| jasond | tspatzier: i think we've discussed it enough. you all raise some good points. we'll try to find an alternate fix. thanks for the discussion | 16:40 |
| tspatzier | jasond: maybe we don't call it fix. Let's come up with a clean solution that also solves similar use cases in the future? | 16:41 |
| jasond | tspatzier: we'll we're in the situation where we're having to support templates with user-data scripts | 16:42 |
| jasond | *well | 16:42 |
| tspatzier | understand. but what I meant with clean solution was a way for have the data you need somewhere where it is intuitive, and then promote this model. | 16:43 |
| tspatzier | btw. shardy: is there some good documentation on how you work with environments, provider templates etc.? I have been asked by colleagues. I think somebody talked about some blogs. But is there something in the docs already? | 16:44 |
| jasond | tspatzier: using environment files is the only option, right? | 16:44 |
| *** vijendar has joined #heat | 16:45 | |
| shardy | tspatzier: I wrote a simple example in a blog post: | 16:45 |
| shardy | http://hardysteven.blogspot.co.uk/2013/10/heat-providersenvironments-101-ive.html | 16:45 |
| shardy | tspatzier: we probably need to improve the actual docs | 16:45 |
| tspatzier | jasond: at least this is what's there today, and I think we should extend this if necessary for your use case. | 16:45 |
| shardy | maybe I'll try to refactor the blog example into a docs howto page or something | 16:45 |
| tspatzier | shardy: thanks! We can think of something for the template guide. There are many folks asking for it, and unless one has followed all past discussions, it is hard to figure out who it works. | 16:46 |
| jasond | tspatzier: i'll mess around and see how it works. will revisit this if necessary. thanks | 16:46 |
| tspatzier | sounds good jasond | 16:46 |
| shardy | tspatzier: There is http://docs.openstack.org/developer/heat/template_guide/environment.html | 16:47 |
| *** radez is now known as radez_g0n3 | 16:47 | |
| shardy | but probably need something a little more introductory in the docs somewhere, and definitely more examples in heat-templates | 16:47 |
| tspatzier | shardy: I know this part. But I guess we need something along a sample use case, how you would split a big template into several provider templates, use them, bind them in the environment etc. | 16:48 |
| *** mestery has quit IRC | 16:49 | |
| lipinski | shardy: did you see the bug I filed yesterday about provider resources with a file:/// reference? I could not get resource_registry to work via (passed) environment file because of that issue. | 16:49 |
| shardy | lipinski: I didn't - IIRC you shoudn't need to pass file:/// as python-heatclient resolves local paths by default | 16:50 |
| lipinski | yes - I could just use the filename without the file:///. But, then it was a relative path and if I was not in the directory of the file(s), it wouldn't work. Anyways, there was another issue preventing me from using (passed) environment files and that is nested resources not working either. | 16:51 |
| lipinski | I had ot switch to a resource_registry in /etc/heat/environment.d to get that to work. | 16:51 |
| shardy | lipinski: you mean propagating the environment to the nested stacks? | 16:51 |
| shardy | That is either by design or a known issue, depending on your perspective | 16:51 |
| shardy | You can pass the parameters from the top level template into each nested resource as a workaround | 16:52 |
| lipinski | shardy: I guess. I have a template with a provider resource. That provider resource refers to another provider resource. The second one wouldn't recognize the resource | 16:52 |
| lipinski | Can you pass a resource registry? | 16:52 |
| shardy | lipinski: Aha, yeah we should fix that, and IMO also provide a way to specify parameters globally in the environment | 16:53 |
| shardy | lipinski: what's the bug #? | 16:53 |
| *** saurabhs has joined #heat | 16:54 | |
| shardy | nvm found it | 16:54 |
| lipinski | shardy: you're faster than me :) | 16:54 |
| *** gokrokve has quit IRC | 16:55 | |
| zaneb | lipinski: so here's the issue... say you register a provider for OS::Nova::Server that contains an OS::Nova::Server | 16:56 |
| lipinski | zaneb: circular? | 16:56 |
| zaneb | if we pass the environment on, it's turtles all the way down | 16:56 |
| zaneb | so that's why we don't | 16:56 |
| zaneb | iirc it's possible to just specify a URL as the resource type though (not using the environment) | 16:57 |
| lipinski | didn't know you could have circular providers like that. (I actually thought I did that accidentally and encountered an engine error) | 16:57 |
| *** chandan_kumar has quit IRC | 16:57 | |
| zaneb | lipinski: well, you can't because we don't pass the environment down to a provider template :) | 16:57 |
| zaneb | but if we did, you could | 16:58 |
| shardy | zaneb: that's a good point, perhaps using anonymous type: foo.yaml providers would work | 16:58 |
| *** harlowja_away is now known as harlowja | 16:58 | |
| lipinski | zaneb: I think circular refernces/providers is another issue. | 16:58 |
| *** gokrokve_ has joined #heat | 16:58 | |
| lipinski | ok - another question then. Can I define a resource_registry in a template (e.g., in my "parent" provider)? | 16:58 |
| zaneb | shardy: ++ | 16:58 |
| *** retr0h has joined #heat | 16:59 | |
| therve | shardy, What do you mean by anonymous types? | 16:59 |
| zaneb | lipinski: no, resource_registry is purely an environment thing | 16:59 |
| zaneb | therve: where the 'type' is just a URL to a provider template | 16:59 |
| shardy | personally I'm not a fan of that interface though, I much prefer the explicit mapping via the type alias in the environment | 17:00 |
| lipinski | zaneb: Then there's no way to have two layers of providers (template -> provider resource -> provider resource) with passed environment | 17:00 |
| therve | So that works already right? | 17:00 |
| zaneb | therve: I believe so | 17:00 |
| shardy | therve: type: foo.yaml instead of type: My::Descriptive::Type | 17:00 |
| therve | Yeah that's what I've been fixing recently :) | 17:00 |
| *** nati_ueno has joined #heat | 17:00 | |
| *** Slower has joined #heat | 17:00 | |
| zaneb | lipinski: not in the environment, but you can do what shardy just said, and specify the provider directly in the (provider) template, rather than the environment, and get two levels of providers | 17:01 |
| harlowja | zaneb but i want my duck | 17:02 |
| harlowja | lol | 17:02 |
| harlowja | why u take away my pet duck | 17:02 |
| zaneb | lol | 17:02 |
| lipinski | zaneb: ok - didn't know that was possible to reference type: <template_url>. Personally, I agree with shardy that that is ugly, but that will work | 17:02 |
| *** tomek_adamczewsk has quit IRC | 17:02 | |
| therve | In practice it's not bad :) | 17:03 |
| lipinski | therve: agreed. Unless you are making a set of templates that are portable. (Depending on how file path resolution is done - which I have yet to try) | 17:03 |
| tspatzier | zaneb: do you have a sec? I need a pointer related to a bug I discussed with shardy earlier: https://bugs.launchpad.net/heat/+bug/1297761 | 17:03 |
| uvirtbot | Launchpad bug 1297761 in heat "No validation of top-level template keys" [Undecided,New] | 17:03 |
| zaneb | tspatzier: sure | 17:04 |
| tspatzier | zaneb: so far we did much of the keyword checking for HOT in the _translate function in HOTemplate, or in the schema classes. | 17:04 |
| tspatzier | _translate does not work for top-level sections since this is called in __geitem__ only. Is there some general validate code as a good entry point? | 17:05 |
| *** wchrisj has quit IRC | 17:05 | |
| *** duncanjw has joined #heat | 17:06 | |
| *** tomek_adamczewsk has joined #heat | 17:06 | |
| zaneb | tspatzier: hmm, I take it you don't want to do it in __init__? | 17:06 |
| tspatzier | zaneb: that was one idea. Like introduce a validate func in the template class and let each sub-class implement it. | 17:06 |
| tspatzier | I would probably do nothing in CfnTemplate, but add the code to fix the bug in HOTemplate. | 17:07 |
| zaneb | I actually think we should fix CfnTemplate too | 17:07 |
| tspatzier | Ok, so we could just check if top-level keys are in SECTIONS | 17:07 |
| zaneb | yep, +1 | 17:07 |
| zaneb | but separate validate() method sounds good | 17:08 |
| *** alexheneveld has quit IRC | 17:08 | |
| zaneb | I would call it from approximately here: https://github.com/openstack/heat/blob/master/heat/engine/parser.py#L338 | 17:08 |
| tspatzier | Ok, let me give it a try. I just wanted to make sure I don't introduce this method if such code is already hidden somewhere else. | 17:08 |
| *** mestery has joined #heat | 17:09 | |
| *** pvaneck has joined #heat | 17:09 | |
| * zaneb wonders if https://review.openstack.org/#/c/82487/ shouldn't be called from there too | 17:09 | |
| *** alexheneveld has joined #heat | 17:09 | |
| tspatzier | zaneb: ok, I would have called the validate() method from __init__ but can also put it there | 17:10 |
| *** sgordon has joined #heat | 17:10 | |
| *** jistr has quit IRC | 17:10 | |
| zaneb | tspatzier: if we put it there it might not break Heat if there are existing stacks in the database that don't pass | 17:11 |
| zaneb | and we can save a database migration to fix them | 17:11 |
| zaneb | maybe. | 17:11 |
| tspatzier | we discussed adding a db migrate step for this. You think we could get around it? | 17:11 |
| *** e0ne_ has quit IRC | 17:15 | |
| *** duncanjw has quit IRC | 17:16 | |
| *** tomek_adamczewsk has quit IRC | 17:17 | |
| *** Slower has quit IRC | 17:19 | |
| *** kgriffs is now known as kgriffs_afk | 17:22 | |
| *** Slower has joined #heat | 17:27 | |
| zaneb | tspatzier: I think so, iirc Stack.validate() only gets called on create and update, it shouldn't get called on existing templates, so we should be able to improve validation without a db migration | 17:27 |
| zaneb | *should* being the operative word ;) you'd have to check | 17:27 |
| tspatzier | zaneb: thanks. I'll try it out :-) | 17:28 |
| *** scollier_ has joined #heat | 17:31 | |
| *** scollier has quit IRC | 17:32 | |
| SpamapS | I filed a blueprint around validation btw.. | 17:36 |
| *** IlyaE has quit IRC | 17:37 | |
| SpamapS | https://blueprints.launchpad.net/heat/+spec/user-friendly-template-errors | 17:38 |
| SpamapS | tspatzier: ^^ | 17:38 |
| SpamapS | Something to consider when talking about validation. If you haven't seen it yet.. our validation errors are.. REALLY painful to actually use. | 17:38 |
| SpamapS | (btw we should never validate things already accepted) | 17:39 |
| SpamapS | crap happens. Just deal with it at runtime for those. | 17:39 |
| tspatzier | SpamapS: thanks for the pointer. I think some of the items are recorded as bugs already. The fix for the bug I am working on will only partly address the bp for now. So overall there will be more work to be done. But I definitely agree this makes sense! | 17:40 |
| *** julienvey has quit IRC | 17:40 | |
| *** gokrokve_ has quit IRC | 17:41 | |
| *** erecio has joined #heat | 17:42 | |
| *** tspatzier has quit IRC | 17:48 | |
| *** saurabhs has quit IRC | 17:50 | |
| *** IlyaE has joined #heat | 17:53 | |
| *** radez_g0n3 is now known as radez | 17:57 | |
| *** saju_m has joined #heat | 17:57 | |
| *** saurabhs has joined #heat | 18:00 | |
| *** lindsayk has joined #heat | 18:01 | |
| openstackgerrit | Steven Hardy proposed a change to openstack/heat: Update heat.conf.sample for keystoneclient 0.7 https://review.openstack.org/83151 | 18:01 |
| shardy | ^^ Looks like a new keystone release just happened and it's breaking us in the gate | 18:01 |
| *** derekh has quit IRC | 18:04 | |
| shardy | hmm, actually that looks wrong | 18:04 |
| * shardy curses at generate_sample.sh | 18:04 | |
| SpamapS | shardy: +2'd.. left a comment, but it seems to me like that is a generated file and thus should be kicked out of the git repo | 18:04 |
| *** lindsayk has quit IRC | 18:05 | |
| shardy | SpamapS: I somewhat agree, other than generate_sample.sh sometimes does some crazy stuff and AFAIK all other projects also check theirs in | 18:05 |
| SpamapS | shardy: o/` dramatic music o/` To the mailing list!! | 18:06 |
| shardy | SpamapS: anything which avoids us breaking every keystoneclient release gets +2 from me.. | 18:06 |
| * SpamapS leaps onto the pole leading to the bat cave.. err.. mailing list | 18:06 | |
| shardy | like, why has it removed clients_trove and replaced it with repeated duplicate options for clients_swift?! | 18:08 |
| *** duncanjw has joined #heat | 18:08 | |
| SpamapS | sent | 18:10 |
| SpamapS | shardy: ruh roh | 18:10 |
| zaneb | shardy: what version of the generate script did you use | 18:12 |
| zaneb | ? | 18:12 |
| shardy | zaneb: the version in heat master | 18:12 |
| *** asalkeld has joined #heat | 18:12 | |
| shardy | so, we're broken by keystoneclient, but have no way to update our config? | 18:13 |
| zaneb | I just left a comment with a link to the bug causing it, but that shouldn't be present in Heat master... | 18:13 |
| zaneb | oh crap | 18:13 |
| shardy | i.e the gate expects whatever wrong stuff generate_sample spits out? | 18:13 |
| zaneb | I know why | 18:13 |
| zaneb | the 'fix' in oslo-incubator that caused it was just a temporary workaround | 18:13 |
| zaneb | the 'real fix' was applied to oslo.config | 18:14 |
| SpamapS | zaneb: see my ML post. This is really silly IMO. | 18:14 |
| zaneb | so any update to olso.config would break us | 18:14 |
| * shardy sighs | 18:14 | |
| shardy | Ok I'll abandon my broken patch and we can work out an alternate plan | 18:14 |
| zaneb | SpamapS: haven't got it yet, but will do | 18:14 |
| SpamapS | shardy: Perhaps we can be be pioneers and remove it from our git tree first. :) | 18:15 |
| * SpamapS completes the drive by and goes to lunch | 18:15 | |
| sdake | zaneb if I run heat-manage db_sync shouldn't it call upgrade every time I run it? | 18:17 |
| sdake | or does it only do so once | 18:17 |
| *** alexheneveld has quit IRC | 18:17 | |
| *** duncanjw has quit IRC | 18:18 | |
| zaneb | sdake: it'll upgrade to the latest version I believe | 18:18 |
| zaneb | but once you're at the latest version, there's no more upgrades to do | 18:19 |
| *** yogesh has joined #heat | 18:19 | |
| sdake | is there any way to get it to un-upgrade | 18:19 |
| zaneb | so to test your migration again, you'd downgrade and then upgrade | 18:19 |
| sdake | how do you downgrade? | 18:19 |
| zaneb | good question, I've never done it with heat-manage | 18:20 |
| sdake | is has no downgrade command | 18:20 |
| sdake | oh well I'll figure out how to test htanks | 18:20 |
| zaneb | sdake: sounds like a feature we could add ;) | 18:20 |
| zaneb | can you not pass a revision to db-sync? | 18:21 |
| sdake | yup thta works thanks | 18:21 |
| sdake | zaneb we are going to have to stick with the whitelist - i think it may hav ebeen possible for people to put non-dates in the version field | 18:23 |
| zaneb | true | 18:24 |
| zaneb | maybe always rewrite non-dates? | 18:24 |
| sdake | ya that could work | 18:24 |
| sdake | Is there something that parses dates besides date()? | 18:25 |
| zaneb | might be something in openstack.common | 18:25 |
| sdake | i'll check that out | 18:25 |
| zaneb | openstack.common.timeutils uses the iso8601 library | 18:28 |
| zaneb | parse_isotime in timeutils might do what you want | 18:28 |
| shardy | zaneb: do you have a handle on what needs to happen re the conf.sample? | 18:28 |
| zaneb | shardy: it's a mess | 18:29 |
| shardy | zaneb: I'm about to stop for some dinner before the meeting and would be good to get the gate unblocked | 18:29 |
| shardy | zaneb: agree | 18:29 |
| zaneb | I put up patches and they got knocked back | 18:29 |
| zaneb | it's hard to find something that works for both us and Nova | 18:29 |
| sdake | zaneb strptime | 18:29 |
| zaneb | sdake: that could also work | 18:30 |
| sdake | parse_isotime includes date and time | 18:30 |
| *** e0ne has joined #heat | 18:30 | |
| shardy | zaneb: Ok, I replied to SpamapS thread saying yes-but-no, I guess we can discuss later | 18:30 |
| shardy | bbiab | 18:30 |
| *** spzala has quit IRC | 18:41 | |
| *** mkollaro has quit IRC | 18:44 | |
| *** andersonvom_ has joined #heat | 18:45 | |
| *** andersonvom has quit IRC | 18:47 | |
| *** mkollaro has joined #heat | 18:48 | |
| *** tango has joined #heat | 18:51 | |
| *** varora has left #heat | 18:55 | |
| *** asalkeld has quit IRC | 18:57 | |
| *** jcru has quit IRC | 19:03 | |
| *** randallburt1 has joined #heat | 19:03 | |
| *** mkollaro has quit IRC | 19:03 | |
| *** randallburt has quit IRC | 19:03 | |
| *** andrew_plunk has quit IRC | 19:03 | |
| *** kgriffs_afk is now known as kgriffs | 19:04 | |
| *** sjmc7 has joined #heat | 19:07 | |
| *** spzala has joined #heat | 19:08 | |
| openstackgerrit | Sergey Kraynev proposed a change to openstack/heat: Initial validation of functions https://review.openstack.org/82486 | 19:09 |
| openstackgerrit | Zane Bitter proposed a change to openstack/heat: Temporary fix for config file generation https://review.openstack.org/83178 | 19:10 |
| openstackgerrit | Sergey Kraynev proposed a change to openstack/heat: Validation functions for resources and outputs https://review.openstack.org/82487 | 19:10 |
| *** wchrisj has joined #heat | 19:14 | |
| *** erecio has quit IRC | 19:16 | |
| openstackgerrit | Sergey Kraynev proposed a change to openstack/heat: Validation functions for resources and outputs https://review.openstack.org/82487 | 19:18 |
| *** andrew_plunk has joined #heat | 19:19 | |
| *** andrew_plunk has quit IRC | 19:25 | |
| *** sabeen has quit IRC | 19:33 | |
| *** spzala has quit IRC | 19:38 | |
| *** andrew_plunk has joined #heat | 19:38 | |
| openstackgerrit | Sergey Kraynev proposed a change to openstack/heat: Adding validation algorithm for get attr functions https://review.openstack.org/82488 | 19:39 |
| *** asalkeld has joined #heat | 19:46 | |
| *** john-n-seattle2 has left #heat | 19:46 | |
| *** rbuilta has quit IRC | 19:47 | |
| *** skraynev_afk is now known as skraynev | 19:54 | |
| skraynev | zaneb: are you here? | 19:54 |
| zaneb | I sure am | 19:54 |
| *** mspreitz has joined #heat | 19:54 | |
| *** tspatzier has joined #heat | 19:55 | |
| skraynev | zaneb: when I validate GetAttr , I use attribute schema | 19:55 |
| zaneb | oh, the meeting is in 5mins? | 19:55 |
| * zaneb has been looking at next week's calendar all week | 19:55 | |
| zaneb | skraynev: right | 19:56 |
| *** bgorski has joined #heat | 19:56 | |
| skraynev | zaneb: but if remember we wants that this function also returnt attribute are not presented in attribute schema, but appeared in property schema | 19:57 |
| zaneb | since when? | 19:57 |
| zaneb | I've never heard of that | 19:57 |
| *** randallburt1 has quit IRC | 19:57 | |
| *** jergerber has joined #heat | 19:58 | |
| skraynev | Hm. in one of review stevebaker said, that we should not duplicate property in attribute schema, because we could get it from properties | 19:59 |
| stevebaker | hmmwha? | 19:59 |
| skraynev | he-he, but it's not possible now | 19:59 |
| zaneb | I'll let stevebaker take this one ;) | 20:00 |
| stevebaker | meeting time | 20:00 |
| zaneb | saved by the bell | 20:00 |
| *** andrew_plunk has quit IRC | 20:01 | |
| skraynev | discuss it later... | 20:02 |
| *** randallburt has joined #heat | 20:03 | |
| *** radez is now known as radez_g0n3 | 20:04 | |
| *** jcru has joined #heat | 20:04 | |
| *** kebray has joined #heat | 20:05 | |
| *** andrew_plunk has joined #heat | 20:05 | |
| sdake_ | stevebaker couple hours until migration patch sorted out | 20:06 |
| *** e0ne has quit IRC | 20:07 | |
| *** saju_m has quit IRC | 20:11 | |
| *** andersonvom_ has quit IRC | 20:22 | |
| *** andersonvom has joined #heat | 20:23 | |
| *** afarrell has joined #heat | 20:25 | |
| *** jprovazn is now known as jprovazn_afk | 20:25 | |
| afarrell | hello - looking to contribute to heat. specifically, tosca translation... | 20:25 |
| *** afarrell_ has joined #heat | 20:27 | |
| asalkeld | hi afarrell | 20:27 |
| asalkeld | we are in #openstack-meeting | 20:27 |
| afarrell | hi - pleased to meet you | 20:27 |
| afarrell | ok, thanks! | 20:28 |
| SpamapS | hm | 20:28 |
| SpamapS | CVE-2014-2525 is _critical_ for Heat users. | 20:28 |
| uvirtbot | SpamapS: ** RESERVED ** This candidate has been reserved by an organization or individual that will use it when announcing a new security problem. When the candidate has been publicized, the details for this candidate will be provided. (http://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2014-2525) | 20:28 |
| asalkeld | libyaml? | 20:29 |
| *** saju_m has joined #heat | 20:30 | |
| zaneb | SpamapS: link? | 20:30 |
| *** BillArnold has joined #heat | 20:33 | |
| *** david-lyle has quit IRC | 20:34 | |
| SpamapS | oh mitre, catch up with the times | 20:34 |
| SpamapS | https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=742732 | 20:35 |
| uvirtbot | Debian bug 742732 in libyaml "libyaml: CVE-2014-2525: input sanitization errors" [Grave,Fixed] | 20:35 |
| SpamapS | asalkeld: correct | 20:35 |
| SpamapS | stevebaker: btw, I made this today.. might be useful to link to it from some software-config docs: https://wiki.openstack.org/wiki/OsCollectConfig | 20:36 |
| *** saju_m has quit IRC | 20:37 | |
| *** afarrell_ has quit IRC | 20:38 | |
| skraynev | zaneb, stevebaker: all is ok, it was bug in my memories. I incorrect rewording stevebakers words about properties. it just was related with that we already know these values and should not present them in attribute schema | 20:40 |
| *** sabeen has joined #heat | 20:44 | |
| *** sabeen has quit IRC | 20:44 | |
| *** sabeen has joined #heat | 20:44 | |
| stevebaker | SpamapS: purdy! | 20:46 |
| *** BillArnold_ has joined #heat | 20:49 | |
| *** BillArnold has quit IRC | 20:49 | |
| *** varora has joined #heat | 20:51 | |
| *** e0ne has joined #heat | 20:57 | |
| *** kfox1111 has joined #heat | 20:58 | |
| tspatzier | afarrell: so you wanted to talk about the TOSCA work? | 21:01 |
| sdake | randallburt your ug was in response to the mistral workflow idea? | 21:01 |
| mspreitz | I also do not like the mistral answer | 21:01 |
| zaneb | randallburt, skraynev: wait, are we talking parameters or properties? | 21:01 |
| skraynev | randallburt: original version is default ? | 21:01 |
| randallburt | sdake: yup | 21:01 |
| shardy | SpamapS, stevebaker: has anyone considered making a resource which runs diskimage-builder? | 21:02 |
| skraynev | zaneb: about properties | 21:02 |
| mspreitz | I also think update should not replace default with previous value | 21:02 |
| * zaneb may have been talking about the wrong thing | 21:02 | |
| randallburt | skraynev: parameters | 21:02 |
| skraynev | he-he | 21:02 |
| afarrell | thomas - just trying to find the window that your message popped up in! | 21:02 |
| zaneb | ok, you two decide between you | 21:02 |
| randallburt | shardy: we have, but its tricky since you'd need a sandboxed environment for building the user's images | 21:02 |
| stevebaker | I was thinking about reimplementing harestarter to "taint" a resource for UpdateReplace then trigger a stack update | 21:02 |
| zaneb | what the question is, then I'll answer it ;) | 21:02 |
| shardy | randallburt: can't you just use a VM as the sandbox and implement it via a provider template? | 21:02 |
| tspatzier | afarrell: What time zone are you in? It's getting late here, so would be good to have a chat tomorrow at a time that works for both of us. | 21:03 |
| shardy | randallburt: then have a URL output as an attribute, which is then passed to an OS::Glance::ImageUploader resource? | 21:03 |
| skraynev | zaneb, randallburt: we could define properties using parameters sections (I think it's possible too) | 21:03 |
| shardy | randallburt: that is the idea I had while discussing ideas with some Tuskar folks last week | 21:03 |
| mspreitz | shardy: interesting discussion around image building | 21:04 |
| skraynev | zaneb, randallburt: but in reality it's two different situations | 21:04 |
| openstackgerrit | Jason Dunsmore proposed a change to openstack/heat: Don't use SSH in Rackspace::Cloud::Server https://review.openstack.org/83218 | 21:04 |
| stevebaker | shardy: I'd like to try implementing a "work" nested stack which deletes the nested stack when it goes to complete, but captures the outputs. You could perform all sorts of compute tasks with that | 21:04 |
| zaneb | skraynev: now I don't even know what you're talking about | 21:04 |
| openstackgerrit | Richard Lee proposed a change to openstack/heat: Add admin_pass as attribute to CloudServer https://review.openstack.org/72745 | 21:04 |
| shardy | randallburt: I'm trying to decide if it's worth hacking out a PoC to see if it works, or if anyone already has | 21:04 |
| randallburt | shardy: sure, but you'd want to then delete that once you've built and moved the image to Glance. will probably take more time than you've saved in the end. | 21:04 |
| openstackgerrit | Richard Lee proposed a change to openstack/heat: Add save_admin_pass property to CloudServer https://review.openstack.org/79481 | 21:04 |
| afarrell | Hi Thomas | 21:04 |
| afarrell | I'm in New Zealand - understand that it might be late in CET! | 21:04 |
| skraynev | zaneb: ok. we have resource with property which has default value [] | 21:05 |
| afarrell | So no worries, I can message later... | 21:05 |
| tspatzier | afarrell: 10am for you now? | 21:05 |
| skraynev | zaneb: create this resource with [val1 , val2] | 21:05 |
| afarrell | Yes that's it | 21:05 |
| shardy | randallburt: the motivation wasn't really time, more definition of the images and usage of the images in one set of templates | 21:05 |
| zaneb | skraynev: ok, it is properties right, not parameters because you mentioned prop_diff before | 21:05 |
| shardy | stevebaker: Interesting idea | 21:05 |
| sdake | re ha restarter if the implementation is wrong lets fix it, if its in the wrong place lets figure out a migration plan | 21:05 |
| zaneb | skraynev: so randallburt is just confusing things ;) | 21:05 |
| skraynev | zaneb: then update this resource using template we this property are not mentioned | 21:05 |
| randallburt | shardy: I don't have any working code around that is worth keeping, so hack on. I think something like this would have lots of good uses if done right. | 21:06 |
| zaneb | skraynev: right, right, I get all that | 21:06 |
| stevebaker | sdake: +1 | 21:06 |
| sdake | zaneb the general "jam it into mistral" seems unsatifying to me because it leaves alot to the imagination | 21:06 |
| randallburt | zaneb, skraynev most likely I am confusing things. | 21:06 |
| afarrell | tspatzier - if you could send email address via linked in, then I will send bio and background info | 21:06 |
| sdake | eg I think it would end up without a real implementation | 21:06 |
| shardy | stevebaker: I was thinking maybe we could do the image building via SoftwareConfig, where you have a SoftwareConfig resource encapsulate the DiB elements, and a deployment hook which runs DiB | 21:06 |
| skraynev | zaneb: =) ok. | 21:06 |
| randallburt | shardy: seems a little overkill to me | 21:06 |
| tspatzier | afarrell: ok, I'll send you a mail and we go from there. Happy to have the discussion :-) | 21:06 |
| shardy | randallburt: but I want to play with all the new toys! ;D | 21:07 |
| stevebaker | shardy: yes, but in a work stack | 21:07 |
| skraynev | zaneb: solution is using new value or default. right? | 21:07 |
| randallburt | shardy: lol, there is that ;) | 21:07 |
| afarrell | tspatizer - cool, thanks. Much appreciated | 21:07 |
| zaneb | skraynev: use the default | 21:07 |
| afarrell | tspatizer - have a good eve :-) | 21:07 |
| *** e0ne has quit IRC | 21:07 | |
| sdake | randallburt does rackspace have any interest in the harestarter functionality being in heat? | 21:08 |
| stevebaker | afarrell: there is this tool too, for tosca -> hot translation http://git.openstack.org/cgit/stackforge/heat-translator/ | 21:08 |
| sdake | or don't really care? | 21:08 |
| skraynev | zaneb: kk | 21:08 |
| mspreitz | re default: REST means use the default | 21:08 |
| shardy | stevebaker: In this case, you couldn't delete the stack until the image has been downloaded, so you'd need a signal-driven decommisioning of the work stack | 21:08 |
| afarrell | stevebaker: thanks for the link. will check it out... | 21:08 |
| randallburt | sdake: I can't speak for "Rackspace", but I personally don't have strong opinions either way | 21:08 |
| zaneb | sdake: HARestarter, and anything else that requires a periodic task, makes scaling heat-engine a pig | 21:08 |
| sdake | i get the periodic task argument | 21:09 |
| zaneb | in fact, it probably makes it impossible at the moment | 21:09 |
| shardy | zaneb: HARestarter doesn't require a periodic task if you use ceilometer for alarms | 21:09 |
| skraynev | zaneb: randallburt gave my other idea: what we should to do if this property has reference on parameter | 21:09 |
| stevebaker | shardy: we would probably need some manner of glance image | 21:09 |
| stevebaker | shardy: glance resource | 21:09 |
| shardy | stevebaker: Yeah I was thinking a simple plugin which takes a URL and uploads it | 21:09 |
| zaneb | skraynev: um, use the parameter value? | 21:10 |
| stevebaker | shardy: yep | 21:10 |
| skraynev | zaneb: yeap. I got answer while describe a situation ;) | 21:10 |
| stevebaker | shardy: streaming the image data in a signal response is probably a non-starter ;) | 21:10 |
| *** tspatzier has quit IRC | 21:11 | |
| *** wchrisj has quit IRC | 21:11 | |
| asalkeld | and i was nono'ed for task like resource ideas:/ | 21:11 |
| asalkeld | glance uploading ... | 21:11 |
| mspreitz | asalkeld: think of it as returning a value via a reference | 21:12 |
| skraynev | zaneb: we have three ways - use corresponding parameter , use default value of parameter if parameter was not passed, get error from client if parameter was not passed and had not default value. | 21:12 |
| shardy | stevebaker: I was suggesting you just expose a URL where the image can be downloaded (run httpd on the vm for example) and point the OS::Glance::Uploader resource at it | 21:12 |
| stevebaker | shardy: yes. uploading to swift and passing a swift url is another option | 21:13 |
| mspreitz | re default: in a REST interface the input is state, not a delta operation. The interpretation should be uniform, not depend on previous state. | 21:13 |
| skraynev | zaneb: anyway thx. for the help. | 21:13 |
| shardy | stevebaker: Ya, probably a few ways to approach it | 21:13 |
| zaneb | image building and HA-restarting are two examples of things that are workflows (verbs, not nouns), and therefore belong in a workflow service, not in Heat. If we had such a service already we would not be thinking about putting them in Heat | 21:13 |
| asalkeld | mspreitz, what ever makes you feel happy with it;) | 21:13 |
| * skraynev go to sleep | 21:13 | |
| skraynev | good night all | 21:14 |
| shardy | stevebaker: Ok, if nobody's already doing it I may take a look into it | 21:14 |
| *** skraynev is now known as skraynev_afk | 21:14 | |
| zaneb | mspreitz: explain PATCH then | 21:14 |
| *** aweiteka has quit IRC | 21:14 | |
| mspreitz | zaneb: OK, I was thinking of PUT | 21:14 |
| zaneb | :) | 21:14 |
| mspreitz | I think UPDATE is more like PUT than PATCH | 21:14 |
| zaneb | yes, and in fact it does use PUT | 21:15 |
| *** afarrell has quit IRC | 21:15 | |
| zaneb | now, if there were an option to do PATCH on an update... | 21:15 |
| stevebaker | we need a PATCH update as well, its not friendly to require restating the template and all parameters | 21:15 |
| zaneb | ++ | 21:15 |
| stevebaker | we have a bug for that | 21:16 |
| mspreitz | Yes I can see that having both is handy. Let's not confuse them into one thing | 21:16 |
| zaneb | mspreitz: +1 | 21:16 |
| *** jprovazn_afk has quit IRC | 21:16 | |
| stevebaker | a separate workflow service could always trigger work items expressed as heat templates | 21:17 |
| mspreitz | Has anybody tried achieving HA restarting via ASG with alarm on population too low? | 21:17 |
| stevebaker | zaneb: ^ | 21:17 |
| shardy | mspreitz: Yes, you can have a heartbeat alarm and trigger scale-up if e.g the instance in the group os powered off or crashes | 21:18 |
| asalkeld | mspreitz, how does it know the health? | 21:18 |
| mspreitz | And: wouldn't that be off by one (or more) if the ASG is not already on top of this? | 21:18 |
| zaneb | stevebaker: yes, in some cases I think that's likely (e.g. triggering stack updates wouldn't be uncommon, most likely) | 21:18 |
| shardy | asalkeld: Is there a ceilo agent now, or do we still bounce stats via heat-api-cloudwatch? | 21:19 |
| mspreitz | The ASG maintains its own "desired count", what HA restarting is not about changing the "desired count" | 21:19 |
| SpamapS | shardy: that would require an API which runs diskimage-builder | 21:19 |
| shardy | last time I checked we still had to use cfn-push-stats | 21:19 |
| *** lifeless has quit IRC | 21:19 | |
| asalkeld | shardy, it bounces | 21:19 |
| SpamapS | shardy: which people have indeed considered | 21:19 |
| shardy | SpamapS: I was suggesting build a VM, install DiB, build image via SoftwareConfig hook, make image available via webserver | 21:19 |
| stevebaker | asalkeld: what does it bounce to? a ceilo API call? | 21:19 |
| shardy | SpamapS: e.g via a provider resource | 21:20 |
| mspreitz | asalkeld: right, ceilo today may count instances but not their health | 21:20 |
| asalkeld | stevebaker, yeah redirects at the api | 21:20 |
| asalkeld | from heat-cfn to ceilo-api | 21:20 |
| SpamapS | shardy: yet another use case for a vm that goes away before the stack reaches COMPLETE | 21:20 |
| mspreitz | shardy: oh, where do I read about heartbeat alarm? | 21:20 |
| shardy | asalkeld: coolm thanks | 21:20 |
| stevebaker | asalkeld: does python-ceilometerclient cli expose those calls? we could call ceilo directly from the server | 21:21 |
| asalkeld | sure | 21:21 |
| asalkeld | just auth | 21:21 |
| SpamapS | shardy: I'd want that image to be pushed into swift though, not made available via webserver | 21:21 |
| asalkeld | you won't want user/pass on the intance | 21:21 |
| sdake | zaneb harestarter as a verb - eg workflow - understand the argument - just need a migration path if it doesn't belong in heat | 21:21 |
| SpamapS | shardy: or even better, glance :) | 21:21 |
| shardy | mspreitz: https://github.com/openstack/heat-templates/blob/master/cfn/F17/WordPress_Single_Instance_With_IHA.template | 21:21 |
| stevebaker | asalkeld: the only hard bit would be getting it working with the heat domain credentials | 21:22 |
| asalkeld | SpamapS, we are building this in solum | 21:22 |
| *** e0ne has joined #heat | 21:22 | |
| SpamapS | shardy: Still there's a pretty good use case there for a dedicated API which just builds images for users. | 21:22 |
| asalkeld | (image customization) | 21:22 |
| SpamapS | asalkeld: Yeah thats what I figured, somebody would build an API for that. | 21:22 |
| SpamapS | though I hope it is based on "at the end the image uploads an image" and not solely on nova snapshots. | 21:22 |
| SpamapS | err | 21:22 |
| SpamapS | at the end the program uploads an image | 21:23 |
| asalkeld | SpamapS, https://github.com/stackforge/solum/blob/master/solum/builder/controllers/v1/image.py | 21:23 |
| zaneb | sdake: the migration path starts with deprecating it so that nobody else gets sucked into using it :) | 21:23 |
| asalkeld | super crude atm | 21:23 |
| shardy | SpamapS: yeah I was thinking expose a URL to avoid putting credentials in the VM, then have a (python) plugin which does the uploading to glance | 21:24 |
| asalkeld | SpamapS, we might move from that to using jenkins/x-workflow thingy to actually do the building | 21:25 |
| *** afarrell has joined #heat | 21:25 | |
| SpamapS | asalkeld: [ insert image builder heat template here ] is how I'd do it. .. too many ways to skin that cat. | 21:26 |
| asalkeld | SpamapS, it needs to be super fast for docker images | 21:26 |
| asalkeld | so don't want to start vms to build | 21:27 |
| sdake | zaneb i see how it starts not how it ends :) | 21:27 |
| SpamapS | asalkeld: right, so your heat template there is Docker::Container ... | 21:27 |
| asalkeld | docker building on docker is *special* | 21:27 |
| asalkeld | need special privileges etc.. | 21:28 |
| zaneb | sdake: a journey of a thousand miles begins with a single step ;) | 21:29 |
| asalkeld | maybe heat could grow something like nodepoop | 21:29 |
| asalkeld | lol | 21:29 |
| mspreitz | shardy: I see how WordPress_Single_Instance_With_IHA.template checks for health. But to respond by hitting a scale-up webhook is off, the problem is not the count in the ASG, the problem is that the count has diverged from useful reality | 21:29 |
| asalkeld | SpamapS, maybe heat could grow something like nodepool | 21:29 |
| asalkeld | and that might be helpful for things like this | 21:30 |
| asalkeld | but again sounds like workflow workers | 21:30 |
| *** tspatzier has joined #heat | 21:30 | |
| *** tspatzier has quit IRC | 21:31 | |
| SpamapS | asalkeld: Actually that's the second use case for node pool like behavior in Heat I"ve heard this week. | 21:31 |
| SpamapS | asalkeld: the other was for LBaaS where they spin up extra LB nodes and don't assign them to users until a previous one goes down for some reason. | 21:31 |
| shardy | mspreitz: No it hasn't, you maintain exactly $n active at any one time | 21:32 |
| asalkeld | yeah we were thinking of it in solum for jenkins workers | 21:32 |
| asalkeld | (as a replacement for nodepool) | 21:32 |
| mspreitz | SpamapS: we are talking about a nodepool resource type, or something built into heat? | 21:32 |
| stevebaker | surely dib isn't safe to run on anything other than a disposable vm | 21:32 |
| mspreitz | shardy: does ASG *maintain* now? | 21:32 |
| SpamapS | mspreitz: just hearing use cases for "pool of similar nodes" in Heat a lot. | 21:32 |
| SpamapS | stevebaker: +1 | 21:32 |
| shardy | mspreitz: I guess we'd need the scale-down parameters to ensure the correct one gets killed on scaledown | 21:33 |
| shardy | mspreitz: there is a bp for that | 21:33 |
| SpamapS | stevebaker: but you can have it un-used in a pool waiting to go.. then build.. then dispose. | 21:33 |
| stevebaker | SpamapS: right, I understand | 21:33 |
| asalkeld | stevebaker, or at least run a small number of jobs | 21:33 |
| asalkeld | and then dispose | 21:33 |
| stevebaker | +1 nodepoop | 21:33 |
| asalkeld | to reduce the exposure | 21:33 |
| asalkeld | ;) | 21:33 |
| *** rpothier_ has quit IRC | 21:34 | |
| SpamapS | asalkeld: I see no difference between 2 and inifinity in that respect.. the 1st job pwns the box, all others are pwned. | 21:34 |
| asalkeld | sure, actually could also use pause/resume here | 21:35 |
| mspreitz | shardy: maybe I got lost. As far as I know, ASG does not *maintain* today. I thought I heard earlier today a suggestion that making an alarm that is triggered by too low working population hit a scale-up policy's webhook could work around lack of maintaining. Now I think we are agreeing it will not, at least without more stuff going on. | 21:35 |
| asalkeld | create 100 vm's and pause them | 21:35 |
| asalkeld | suspend | 21:36 |
| mspreitz | If we could make a composite policy that scales down by removing the dead/broken members and then scales up again, that would be a valid response | 21:37 |
| shardy | mspreitz: well right now, you can define an alarm which will scale up if there is too low working population, and it will scale back down if the broken node comes back online | 21:37 |
| shardy | mspreitz: but there are several other (possibly better) ways to satisfy the HARestarter use-case, some of which I mentioned earlier | 21:38 |
| mspreitz | shardy: that was in the meeting, right? | 21:39 |
| shardy | mspreitz: yes, stuff like triggering a node rebuild via an alarm, or doing a signal initiated resource converge | 21:40 |
| shardy | neither of which are possible today | 21:40 |
| mspreitz | shardy: thanks. Gotta go now. | 21:40 |
| *** asalkeld is now known as asalkeld_bbq | 21:44 | |
| asalkeld_bbq | later food time | 21:45 |
| radix | huh I did not notice to_hot | 21:50 |
| *** randallburt has quit IRC | 21:51 | |
| *** andersonvom has quit IRC | 21:53 | |
| radix | shardy: a resource that runs diskimage-builder is a pretty cool idea | 21:54 |
| *** andersonvom has joined #heat | 21:54 | |
| radix | so is one that boots an instance, runs some config, and saves the result as an image | 21:54 |
| radix | mspreitz: "maintain" is probably going to get a lot of attention next cycle, there are several convergence/error correcting blueprints | 21:57 |
| radix | there's still some question about how the error states will actually be detected | 21:57 |
| *** kgriffs_afk has joined #heat | 21:59 | |
| *** mspreitz has quit IRC | 21:59 | |
| *** kgriffs_afk has left #heat | 22:00 | |
| *** afarrell has quit IRC | 22:02 | |
| *** afarrell_ has joined #heat | 22:02 | |
| *** blamar has quit IRC | 22:03 | |
| *** andersonvom has quit IRC | 22:09 | |
| *** rcleere has quit IRC | 22:09 | |
| *** achampion has quit IRC | 22:19 | |
| *** vijendar has quit IRC | 22:23 | |
| *** yogesh has quit IRC | 22:26 | |
| kfox1111 | Is there a way to respond to a review? | 22:28 |
| radix | kfox1111: kinda? you have to add your own review | 22:28 |
| *** jergerber has quit IRC | 22:28 | |
| kfox1111 | https://review.openstack.org/#/c/82903/ Thomas left me a review, and the second part says the templates are not hot, but I ran the autoconvert tool on it. Not sure whats not hot about it. | 22:30 |
| *** andrew_plunk has quit IRC | 22:31 | |
| kfox1111 | I'm going to post a new version with the documentation change, but don't want to loose track of the other comment. | 22:31 |
| stevebaker | kfox1111: just make sure you keep the same Change-Id in the commit comment | 22:32 |
| *** jmckind has quit IRC | 22:37 | |
| kfox1111 | ah. ok. thanks. | 22:38 |
| *** blamar has joined #heat | 22:42 | |
| kfox1111 | So, when I ammend the commit, I delete the old commit message and replace it with the new comments, along with keeping the change-id? | 22:44 |
| radix | kfox1111: not sure what you mean about replacing it with the new comments | 22:44 |
| radix | yes, you can update the commit message, just leave the change-id | 22:44 |
| kfox1111 | or should I leave the commit message alone, commit the new change, then review it to make the comment? | 22:45 |
| kfox1111 | Ah, yeah. I think I see how it works. the comment goes in the top right of the web page. The reviews show up below. | 22:46 |
| *** IlyaE has quit IRC | 22:47 | |
| openstackgerrit | Kevin Fox proposed a change to openstack/heat-templates: Add a scalable ElasticSearch cluster set of templates in Hot format. https://review.openstack.org/82903 | 22:48 |
| *** e0ne has quit IRC | 22:49 | |
| stevebaker | SpamapS: I'm trying to figure out how to use the new split out deployments json files | 22:52 |
| *** blamar has quit IRC | 22:53 | |
| stevebaker | SpamapS: before I could just have an oac template {{deployments}} | 22:53 |
| SpamapS | stevebaker: they're merged | 22:54 |
| SpamapS | stevebaker: trouble with {{deployments}} is that now you end up having to prefix _everything_ with {{deployments.deploymentname.....}} and deploymentname is not necessarily going to be known | 22:54 |
| stevebaker | SpamapS: but other hooks need the inputs and outputs, thats all gone | 22:54 |
| SpamapS | stevebaker: it isn't gone, or at least, it shouldn't be | 22:55 |
| SpamapS | stevebaker: If it is, file a bug, I meant to leave it behind. Entirely possible I didn't. :P | 22:55 |
| stevebaker | ok | 22:56 |
| SpamapS | stevebaker: the config should be gone, but not the inputs. | 22:56 |
| SpamapS | as in, the config part should go into its own file, but deployments and inputs should stay behind | 22:56 |
| stevebaker | SpamapS: where would the inputs be? cfn.json? | 22:56 |
| SpamapS | stevebaker: right | 22:56 |
| SpamapS | stevebaker: though IMO, if you want access.. put it in config.. ;) | 22:57 |
| SpamapS | which may be why that use case was lost.. because in my head.. I'm like "no.. don't do that." | 22:57 |
| *** duncanjw has joined #heat | 22:57 | |
| stevebaker | I probably need to do something else for my {{deployments}} template anyway. oac gives it read-everyone permissions | 22:58 |
| *** harlowja has quit IRC | 22:58 | |
| *** blomquisg has quit IRC | 23:00 | |
| openstackgerrit | Richard Lee proposed a change to openstack/heat: Add save_admin_pass property to CloudServer https://review.openstack.org/79481 | 23:00 |
| openstackgerrit | Richard Lee proposed a change to openstack/heat: Add admin_pass as attribute to CloudServer https://review.openstack.org/72745 | 23:00 |
| *** arbylee has quit IRC | 23:00 | |
| *** afarrell_ has quit IRC | 23:03 | |
| *** harlowja has joined #heat | 23:03 | |
| *** duncanjw has quit IRC | 23:06 | |
| *** afarrell has joined #heat | 23:07 | |
| SpamapS | stevebaker: oac will keep existing perms | 23:07 |
| SpamapS | stevebaker: so install -m 0640 -o root -g root /dev/null /your/dest/file before hand works | 23:07 |
| *** bgorski has quit IRC | 23:07 | |
| *** Trozz has joined #heat | 23:10 | |
| *** sgordon has quit IRC | 23:10 | |
| *** dims_ has quit IRC | 23:12 | |
| *** kebray has quit IRC | 23:13 | |
| *** blomquisg has joined #heat | 23:14 | |
| *** IlyaE has joined #heat | 23:14 | |
| *** afarrell has quit IRC | 23:15 | |
| openstackgerrit | Steven Dake proposed a change to openstack/heat: Migrate invalid template versions in database https://review.openstack.org/83241 | 23:15 |
| openstackgerrit | Zane Bitter proposed a change to openstack/heat: Update heat.conf.sample https://review.openstack.org/83242 | 23:15 |
| stevebaker | SpamapS: ok, use install to install my oac template? | 23:16 |
| SpamapS | stevebaker: no | 23:17 |
| SpamapS | stevebaker: the destination file | 23:17 |
| SpamapS | stevebaker: if it already exists, oac will copy its perms/owner to the new version | 23:17 |
| stevebaker | ooooh | 23:17 |
| stevebaker | an empty file | 23:17 |
| SpamapS | right | 23:17 |
| *** pablosan is now known as ZZpablosan | 23:18 | |
| SpamapS | you can touch and chmod too.. but I'm an install fan myself ;) | 23:18 |
| *** duncanjw has joined #heat | 23:18 | |
| stevebaker | right | 23:18 |
| SpamapS | stevebaker: if you go by the tests I added for deployments support, we left whether or not deployments still exists in 'cfn' undefined. | 23:20 |
| stevebaker | SpamapS: I think I have a fix. do I need to raise a bug? | 23:20 |
| SpamapS | stevebaker: but in the code, I see we delete it from the original. | 23:20 |
| SpamapS | stevebaker: no if you have a fix, just fire away | 23:21 |
| stevebaker | righto | 23:21 |
| SpamapS | stevebaker: please do make sure we test that though.. so we don't accidentally destroy it again ;) | 23:21 |
| stevebaker | yep | 23:21 |
| SpamapS | stevebaker: I wish all pieces of tripleO were as small as os-collect-config sometimes. :-P | 23:22 |
| *** nati_ueno has quit IRC | 23:23 | |
| stevebaker | its about to get bigger, I'll write a deployment api collector after this | 23:23 |
| *** nati_ueno has joined #heat | 23:23 | |
| SpamapS | stevebaker: I actually wondered about separating this into a new thing, instead of a collector, call it a distributor. | 23:24 |
| SpamapS | stevebaker: that way if we encounter it in the heat native api (instead of cfn) we'd just run through the same code path. | 23:24 |
| *** faramir1 has joined #heat | 23:25 | |
| stevebaker | that makes sense | 23:26 |
| *** asalkeld_bbq has quit IRC | 23:26 | |
| *** alexpilotti has quit IRC | 23:28 | |
| *** duncanjw has quit IRC | 23:29 | |
| *** dims_ has joined #heat | 23:30 | |
| *** sjmc7 has quit IRC | 23:32 | |
| *** IlyaE has quit IRC | 23:37 | |
| *** sabeen has quit IRC | 23:37 | |
| *** duncanjw has joined #heat | 23:38 | |
| *** saurabhs has quit IRC | 23:39 | |
| *** Trozz has quit IRC | 23:40 | |
| *** ZZpablosan is now known as pablosan | 23:43 | |
| *** afarrell has joined #heat | 23:50 | |
| stevebaker | SpamapS: another thing, you're writing out .json files but much config will not be json syntax | 23:51 |
| *** QinJin has joined #heat | 23:56 | |
| *** Victor_HU_ has joined #heat | 23:59 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!