| *** openstack has joined #heat | 13:49 | |
| *** robsparker has joined #heat | 13:50 | |
| *** vijendar has quit IRC | 13:50 | |
| *** vijendar has joined #heat | 13:50 | |
| *** edmund has joined #heat | 13:53 | |
| jdandrea | therve: One more clarification question. I may subclass StackResource with a (relatively) clear conscience, but subclassing any of the built-in heat.engine.resources carries risk of my subclasses breaking, going forward, is that correct? | 13:55 |
|---|---|---|
| *** aweiteka has quit IRC | 13:55 | |
| *** ramishra has joined #heat | 13:59 | |
| *** arbylee has joined #heat | 13:59 | |
| *** bvivek has joined #heat | 14:04 | |
| jdandrea | (... because StackResource is abstract and the others are concrete, which might be at the heart of the matter.) | 14:08 |
| *** aweiteka has joined #heat | 14:08 | |
| *** swygue has quit IRC | 14:08 | |
| *** sabeen has quit IRC | 14:13 | |
| *** sabeen has joined #heat | 14:14 | |
| *** rwsu has joined #heat | 14:16 | |
| *** ckmvishnu has quit IRC | 14:17 | |
| *** DandyPandy has joined #heat | 14:21 | |
| *** FL1SK has joined #heat | 14:22 | |
| therve | jdandrea, That's my opinion at least. | 14:23 |
| jdandrea | therve: Got it. Thank you! Appreciated. | 14:23 |
| *** radez is now known as radez_g0n3 | 14:24 | |
| *** aweiteka has quit IRC | 14:25 | |
| *** radez_g0n3 is now known as radez | 14:27 | |
| openstackgerrit | unmesh-gurjar proposed a change to openstack/python-heatclient: Added timeout and rollback params in stack-preview https://review.openstack.org/103552 | 14:28 |
| *** nkhare has joined #heat | 14:31 | |
| *** arunrajan has joined #heat | 14:36 | |
| *** jdandrea has quit IRC | 14:38 | |
| *** jdandrea has joined #heat | 14:40 | |
| *** ericgoncz has joined #heat | 14:44 | |
| zaneb | jdandrea: yeah, I would agree with therve. you can subclass an internal resource, but if you start overriding/relying on parts that are not part of the public plugin interface then there's a good chance of them breaking in the future | 14:45 |
| zaneb | fwiw I wrote that wiki page, so you can blame me | 14:46 |
| *** jprovazn has quit IRC | 14:46 | |
| * zaneb runs | 14:46 | |
| jdandrea | zaneb: :) | 14:46 |
| *** kopparam_ has quit IRC | 14:47 | |
| *** kopparam has joined #heat | 14:48 | |
| jdandrea | zaneb: Thanks! I agree, I do not wish to override parts that aren't part of the public interface. In the case of NestedStack, for instance, I want to alter how child_template's urlfetch behaves, and that could lead to bad times on my end. | 14:50 |
| zaneb | jdandrea: so handle_create() just calls child_template() and then create_with_template() | 14:52 |
| *** kopparam has quit IRC | 14:52 | |
| zaneb | handle_create() and create_with_template() are both parts of the public API | 14:52 |
| * jdandrea nods | 14:52 | |
| jdandrea | That is where I'd need to introduce things. | 14:52 |
| zaneb | (of Resource and StackResource respectively) | 14:52 |
| jdandrea | Not at child_template. | 14:53 |
| zaneb | so you'd probably be better to override handle_create() than to override child_template() | 14:53 |
| *** maxskew_ has quit IRC | 14:53 | |
| *** bvivek has quit IRC | 14:53 | |
| zaneb | I'd expect that if you did that the chances of it breaking would go from low to very low | 14:54 |
| jdandrea | Hehe. :) | 14:54 |
| jdandrea | Yes. In the process, then, I'd be duplicating what's in _create_with_template and child_template, but with one change (in the call to urlfetch). However, it keeps within the public interface, which is good! | 14:54 |
| *** maxskew has joined #heat | 14:54 | |
| jdandrea | I'm good with that. | 14:54 |
| zaneb | though we're not guaranteeing that that class will always exist at heat.engine.resources.stack.NestedStack | 14:55 |
| zaneb | e.g. it may move | 14:55 |
| *** zhiyan is now known as zhiyan_ | 14:55 | |
| jdandrea | Ah, got it. | 14:55 |
| zaneb | so you will always have some work to do on upgrades to verify that plugins haven't broken | 14:55 |
| zaneb | as long as you're aware of that, it's probably not a big problem | 14:56 |
| jdandrea | Indeed. I'm aware of and accept that. I appreciate that clarification. | 14:56 |
| *** ericgoncz has quit IRC | 14:56 | |
| *** zhiyan_ is now known as zhiyan | 14:56 | |
| jdandrea | Of course, I'd like to minimize/mitigate, and sticking to the public interface is a good way to do that. | 14:57 |
| *** ericgoncz has joined #heat | 14:57 | |
| zaneb | yeah, so as therve said, StackResource is the true public interface | 14:57 |
| jdandrea | Excellent. That's the (abstract) one I want to work off of. | 14:58 |
| zaneb | it's up to you how much risk you want to take ;) | 14:58 |
| jdandrea | ;) | 14:58 |
| jdandrea | Speaking of risk ... | 14:58 |
| *** zhiyan is now known as zhiyan_ | 14:59 | |
| *** sballe has joined #heat | 14:59 | |
| jdandrea | If I wanted to take something like, say, OS::Cinder::VolumeAttachment, and give it additional properties and attributes (nothing that volume.py would know about though, just to give it the appearance of a single resource "with extra features"), would the same cautions apply? That is, could I subclass it and stick to the public interface with the same (low, but potential) risks? | 15:00 |
| jdandrea | (I would give it a different name as well.) | 15:01 |
| *** aweiteka has joined #heat | 15:01 | |
| zaneb | tbh I think if you tried it in that case you would end up rewriting almost everything anyway | 15:01 |
| jdandrea | Oh! | 15:02 |
| jdandrea | Hmm. I probably don't want to do that. Perhaps I would encapsulate the VolumeAttachment resource within my plugin (if that's possible). | 15:02 |
| zaneb | the reason it's easy in the NestedStack case is that all of the heavy lifting is already done by StackResource, which is public | 15:02 |
| * jdandrea nods | 15:02 | |
| *** FrancoisBillant has joined #heat | 15:03 | |
| zaneb | again, you can do it, it will work, it will *probably* still work tomorrow, but you'll have to be vigilant for changes and plan to reverify whenever you upgrade | 15:04 |
| *** nkhare has quit IRC | 15:04 | |
| jdandrea | Eternal vigilance! Yes, understood and accepted. | 15:05 |
| *** tomek_adamczewsk has quit IRC | 15:05 | |
| shardy | jdandrea: If all you wanted was to combine resources and expose them as a new resource type, you can use nested stack templates exposed as a provider resource | 15:09 |
| *** aweiteka has quit IRC | 15:09 | |
| shardy | jdandrea: the complexity is when you require logic not expressible via a heat template, then you have to use a resource plugin | 15:10 |
| jdandrea | shardy: Exactly, and that is, in fact, part of it. StackResource aside, there is one other use case where a plugin is necessary. | 15:10 |
| jdandrea | ... | 15:10 |
| jdandrea | Suppose I have OS::Cinder::VolumeAttachment. I want to execute some extra logic before that resource is created, perhaps altering that resource's parameters along the way. ... | 15:11 |
| *** nati_ueno has joined #heat | 15:12 | |
| jdandrea | My initial thought was: "Oh! I'll make My::Own::Plugin and have it subclass VolumeAttachment." I would stick to the public interface, ofc, but I would add whatever new parameters I needed (whoops - naming collision risk in the future) ..." | 15:12 |
| shardy | jdandrea: it all depends on what that "extra logic" actually does, in some cases you can just have a nested stack which does $stuff, whose outputs are then used as properties to some other resource | 15:13 |
| jdandrea | Plan B: "I'll make My::Own::Plugin and encapsulate a VolumeAttachment resource inside of that ... somehow ... and then I manage all of the parameters, passing things in to VA as needed." | 15:13 |
| shardy | jdandrea: If whatever you want to encapsulate is expressible somehow via a heat template (even if, say, you spin up an instance to do the special logic), then you can do this: | 15:15 |
| shardy | https://github.com/hardys/demo_templates/blob/master/juno_summit_intro_to_heat/example4_provider_environment/server_with_volume_env.yaml#L22 | 15:15 |
| shardy | https://github.com/hardys/demo_templates/blob/master/juno_summit_intro_to_heat/example4_provider_environment/env_server_with_volume.yaml#L8 | 15:15 |
| *** ifarkas_ has quit IRC | 15:16 | |
| shardy | So My::Own::Plugin can just refer to a heat stack template, instead of an actual plugin, with all the complexity implied by that | 15:16 |
| *** rbuilta has joined #heat | 15:16 | |
| shardy | obviously it doesn't work for all use-cases, but it's something to consider before implementing a plugin, because the maintenance overhead is probably much lower | 15:16 |
| jdandrea | shardy: Ah. I've looked at using params and the registry before ... but it didn't seem to fit what I wanted to do. There may just be a disconnect in how I'm understanding it. | 15:17 |
| jdandrea | I know we'd like to avoid spinning up an instance to do the special logic if we can help it. (We will call out to another service running on the same server where Heat is, in fact.) | 15:18 |
| *** chandan_kumar has quit IRC | 15:18 | |
| *** david-lyle has joined #heat | 15:18 | |
| * jdandrea studies it again | 15:18 | |
| *** jcoufal has quit IRC | 15:18 | |
| *** arunrajan has quit IRC | 15:19 | |
| *** lazy_prince is now known as killer_prince | 15:20 | |
| *** arbylee has quit IRC | 15:20 | |
| jdandrea | Ok. What I like about this example is it lets you define properties for My::Server::WithVolume such that it looks like any other resource, but under the hood it's really parceling those parameters out to other resources (a server and a volume), which is great! | 15:22 |
| *** erecio has quit IRC | 15:22 | |
| shardy | jdandrea: yup, that's basically the whole cool thing about provider resources :) | 15:22 |
| jdandrea | :) | 15:22 |
| *** andreaf_ has quit IRC | 15:23 | |
| jdandrea | So my question (for me) becomes ... can I use that pattern to do what I'm looking to do. My initial answer to that was "No." I'm revisiting it. | 15:23 |
| *** kebray has joined #heat | 15:25 | |
| *** ckmvishnu has joined #heat | 15:25 | |
| *** arbylee has joined #heat | 15:25 | |
| jdandrea | Let's see. In the yaml that I include, that is my parallel to server_with_volume.yaml in that example ... if I had OS::Cinder::VolumeAttachment in there, plus some *other* resource (a plugin of my design) that VA depended upon, could I reference attributes of my resource when setting properties of the VA resource? That is, can I use get_attr outside of the outputs section in a template? | 15:27 |
| shardy | jdandrea: Yes, you can use get_attr in the resources or outputs section | 15:28 |
| jdandrea | :-D | 15:28 |
| jdandrea | Ah! That was my disconnect. That is *very* good news! | 15:28 |
| shardy | jdandrea: and the outputs of your nested template, become attributes of the resource alias/name you define in the environment | 15:29 |
| jdandrea | That means, instead of subclassing, I could use a provider resource that encapsulates both the built-in resource and my new resource, make the built-in depend upon the new one, and have the results just flow right in. | 15:29 |
| *** ckmvishnu_ has joined #heat | 15:29 | |
| jdandrea | shardy: Yes! This is true (and also good). | 15:29 |
| jdandrea | I'm starting to like this approach more. | 15:29 |
| *** htruta has quit IRC | 15:30 | |
| *** htruta has joined #heat | 15:30 | |
| *** ckmvishnu_ has left #heat | 15:31 | |
| jdandrea | That insight is a big help. Thank you! I am going to give this a try now and see how it flies. | 15:32 |
| *** rakesh_hs has quit IRC | 15:32 | |
| *** ckmvishnu has quit IRC | 15:33 | |
| shardy | jdandrea: np, let us know how it goes :) | 15:34 |
| shardy | jdandrea: note, you'll need to use the heatclient CLI interface to heat, as the environment/nested stack interfaces via horizon are not quite done yet | 15:35 |
| shardy | ref bug 1322258 | 15:35 |
| uvirtbot | Launchpad bug 1322258 in horizon "Heat environments don't work for local files" [High,In progress] https://launchpad.net/bugs/1322258 | 15:35 |
| *** achampio1 has joined #heat | 15:35 | |
| jdandrea | shardy: Thanks! I will. Oh yes, we're CLI for this all the way. | 15:35 |
| *** achampion has quit IRC | 15:37 | |
| *** dsneddon has joined #heat | 15:43 | |
| *** beekneemech is now known as bnemec | 15:45 | |
| *** nati_ueno has quit IRC | 15:47 | |
| *** pasquier-s has quit IRC | 15:55 | |
| *** aweiteka has joined #heat | 15:55 | |
| *** chandan_kumar has joined #heat | 15:59 | |
| *** blamar has joined #heat | 16:00 | |
| *** julienvey has quit IRC | 16:01 | |
| *** arunrajan has joined #heat | 16:03 | |
| *** julienvey has joined #heat | 16:05 | |
| *** erecio has joined #heat | 16:11 | |
| *** shakamunyi has joined #heat | 16:15 | |
| *** shakamunyi has quit IRC | 16:15 | |
| *** shakamunyi has joined #heat | 16:16 | |
| *** shakamunyi has quit IRC | 16:17 | |
| *** shakamunyi has joined #heat | 16:17 | |
| *** shakamunyi has quit IRC | 16:18 | |
| *** shakamunyi has joined #heat | 16:18 | |
| openstackgerrit | Thomas Spatzier proposed a change to openstack/heat-specs: Action-aware software config specification https://review.openstack.org/98742 | 16:25 |
| *** andreaf_ has joined #heat | 16:30 | |
| *** morganfainberg_Z is now known as morganfainberg | 16:30 | |
| *** ccrouch has joined #heat | 16:31 | |
| *** arunrajan has quit IRC | 16:32 | |
| *** arunrajan has joined #heat | 16:36 | |
| *** arunrajan has joined #heat | 16:36 | |
| *** achampion has joined #heat | 16:40 | |
| *** mohits has joined #heat | 16:40 | |
| *** andreaf_ has quit IRC | 16:41 | |
| *** DandyPandy has quit IRC | 16:42 | |
| *** achampio1 has quit IRC | 16:42 | |
| *** ramishra has quit IRC | 16:45 | |
| *** ramishra has joined #heat | 16:45 | |
| *** daneyon has joined #heat | 16:47 | |
| *** daneyon has quit IRC | 16:50 | |
| *** ramishra has quit IRC | 16:50 | |
| *** daneyon has joined #heat | 16:50 | |
| *** arunrajan has quit IRC | 16:53 | |
| *** slick666 has joined #heat | 16:56 | |
| *** yogeshmehra has joined #heat | 16:57 | |
| *** SpamapS_ is now known as SpamapS | 16:57 | |
| *** SpamapS has quit IRC | 16:57 | |
| *** SpamapS has joined #heat | 16:57 | |
| *** bvivek has joined #heat | 17:03 | |
| *** achampio1 has joined #heat | 17:03 | |
| *** derekh_ has quit IRC | 17:05 | |
| *** achampion has quit IRC | 17:06 | |
| *** jdandrea has quit IRC | 17:08 | |
| *** nati_ueno has joined #heat | 17:14 | |
| *** gokrokve has joined #heat | 17:15 | |
| *** harlowja has joined #heat | 17:17 | |
| *** alexheneveld has quit IRC | 17:21 | |
| *** kopparam has joined #heat | 17:22 | |
| *** kopparam has quit IRC | 17:26 | |
| *** arbylee has quit IRC | 17:30 | |
| *** arbylee has joined #heat | 17:30 | |
| *** danielbruno has joined #heat | 17:32 | |
| *** bvivek has quit IRC | 17:41 | |
| *** DandyPandy has joined #heat | 17:44 | |
| *** DandyPandy has quit IRC | 17:45 | |
| *** DandyPandy has joined #heat | 17:46 | |
| *** edmund has quit IRC | 17:48 | |
| *** julienvey has quit IRC | 17:49 | |
| *** tspatzier has quit IRC | 17:50 | |
| *** erecio has quit IRC | 17:54 | |
| *** andersonvom has quit IRC | 17:54 | |
| *** erecio has joined #heat | 17:55 | |
| *** mohits has quit IRC | 17:55 | |
| *** edmund has joined #heat | 17:56 | |
| *** swygue has joined #heat | 17:56 | |
| *** daneyon has quit IRC | 17:56 | |
| *** arbylee has quit IRC | 17:58 | |
| *** Qiming has quit IRC | 17:59 | |
| *** kebray has quit IRC | 18:01 | |
| *** ericgoncz has quit IRC | 18:03 | |
| *** ericgoncz has joined #heat | 18:05 | |
| *** asalkeld has joined #heat | 18:08 | |
| *** swygue has quit IRC | 18:09 | |
| *** daneyon has joined #heat | 18:09 | |
| *** daneyon has quit IRC | 18:10 | |
| *** achampion has joined #heat | 18:11 | |
| *** daneyon has joined #heat | 18:11 | |
| *** achampio1 has quit IRC | 18:13 | |
| *** jdandrea has joined #heat | 18:13 | |
| *** HPJL has joined #heat | 18:15 | |
| *** yogeshmehra has quit IRC | 18:17 | |
| *** andreaf_ has joined #heat | 18:20 | |
| *** yogeshme_ has joined #heat | 18:22 | |
| *** andreaf_ has quit IRC | 18:23 | |
| *** bgorski has quit IRC | 18:24 | |
| *** daneyon has quit IRC | 18:25 | |
| *** andreaf_ has joined #heat | 18:27 | |
| *** IlyaE has joined #heat | 18:31 | |
| *** lsmola__ has joined #heat | 18:31 | |
| *** IlyaE has quit IRC | 18:35 | |
| *** IlyaE has joined #heat | 18:37 | |
| *** JR___ has joined #heat | 18:40 | |
| *** IlyaE has quit IRC | 18:40 | |
| *** tango has joined #heat | 18:40 | |
| *** DandyPandy_ has joined #heat | 18:40 | |
| *** DandyPandy has quit IRC | 18:41 | |
| *** shadower has quit IRC | 18:41 | |
| *** jprovazn has joined #heat | 18:42 | |
| *** shadower has joined #heat | 18:42 | |
| zaneb | jdandrea: btw if you use get_attr in another resource, Heat creates a dependency automatically, you don't even have to specify that separately :) | 18:44 |
| *** IlyaE has joined #heat | 18:45 | |
| *** swygue has joined #heat | 18:47 | |
| *** rbuilta has quit IRC | 18:48 | |
| *** lsmola__ has quit IRC | 18:50 | |
| jdandrea | zaneb: I like that! | 18:50 |
| *** swygue has quit IRC | 18:51 | |
| *** andreaf has quit IRC | 18:51 | |
| *** andreaf has joined #heat | 18:52 | |
| *** randallburt has joined #heat | 18:52 | |
| *** randallburt has quit IRC | 18:52 | |
| *** randallburt has joined #heat | 18:52 | |
| *** swygue has joined #heat | 18:53 | |
| jdandrea | So far, so good (with the provider resource). | 18:54 |
| jdandrea | As for the subclassing of the abstract StackResource class ... I can get at the parsed template's attributes using Outputs.name-of-attr. That is also good. | 18:54 |
| *** arbylee has joined #heat | 18:55 | |
| jdandrea | However (and this may be something remedial that I'm missing), if I want to squirrel away some info during, say, handle_create ... and I put that in self.some_var, it ends up being empty when accessed from within my _resolve_attribute (or FnGetAtt as the case may be). | 18:56 |
| *** arbylee has quit IRC | 18:56 | |
| *** arbylee has joined #heat | 18:56 | |
| *** rbuilta has joined #heat | 18:58 | |
| *** arunrajan has joined #heat | 18:59 | |
| randallburt | jdandrea: if you need to have access to something across object instantiations, you may need to save it in resource_data. | 19:00 |
| jdandrea | Ah, thank you. I was thinking there might need to be some persistence under the hood (still discovering all of that). I will look through the resources for an example of that. | 19:01 |
| jdandrea | So self.data_set(name, value, redact) and self.data().get(name). | 19:02 |
| jpeeler | zaneb: is there a way in the userdata to escape the text so it isn't intrepreted? i'm trying to use a here doc, but the variables and command substitution is getting evaluated before being put in the file. | 19:02 |
| *** lsmola__ has joined #heat | 19:04 | |
| *** SnowDust has joined #heat | 19:04 | |
| jdandrea | randallburt: That worked - thank you! | 19:06 |
| randallburt | jdandrea: np | 19:06 |
| *** adam_g` is now known as adam_g | 19:07 | |
| SpamapS | jpeeler: have you considered just putting the script in swift, and fetching it? | 19:07 |
| jpeeler | SpamapS: i have considered fetching it, maybe not from swift. But it's like 8 lines. | 19:08 |
| *** dims has quit IRC | 19:08 | |
| SpamapS | jpeeler: or have yu tried using get_file ? | 19:08 |
| SpamapS | you even :p | 19:08 |
| *** lsmola__ has quit IRC | 19:08 | |
| jpeeler | SpamapS: that's probably a good thing to look, let me do that | 19:10 |
| *** nati_ueno has quit IRC | 19:11 | |
| *** nati_ueno has joined #heat | 19:12 | |
| *** arunrajan has quit IRC | 19:13 | |
| *** bnemec has quit IRC | 19:18 | |
| *** swygue has quit IRC | 19:20 | |
| *** bnemec has joined #heat | 19:21 | |
| *** ericgoncz has quit IRC | 19:27 | |
| *** IlyaE has quit IRC | 19:27 | |
| *** ericgoncz has joined #heat | 19:28 | |
| *** andersonvom has joined #heat | 19:32 | |
| *** IlyaE has joined #heat | 19:34 | |
| *** sileht_ has quit IRC | 19:34 | |
| *** sileht has joined #heat | 19:35 | |
| *** mspreitz has joined #heat | 19:35 | |
| *** arunrajan has joined #heat | 19:35 | |
| *** arunrajan has quit IRC | 19:38 | |
| *** arunrajan has joined #heat | 19:38 | |
| *** daneyon has joined #heat | 19:39 | |
| *** alexheneveld has joined #heat | 19:48 | |
| *** kebray has joined #heat | 19:49 | |
| *** chandan_kumar has quit IRC | 19:52 | |
| *** shakamunyi has quit IRC | 19:52 | |
| *** FrancoisBillant has quit IRC | 19:52 | |
| *** shakamunyi has joined #heat | 19:53 | |
| *** shakamunyi has quit IRC | 19:53 | |
| *** nati_ueno has quit IRC | 19:55 | |
| *** nati_ueno has joined #heat | 19:55 | |
| *** dims has joined #heat | 19:56 | |
| *** asalkeld has quit IRC | 20:00 | |
| *** rbuilta has quit IRC | 20:01 | |
| *** dsneddon has quit IRC | 20:04 | |
| *** pafuent has left #heat | 20:05 | |
| *** arunrajan has quit IRC | 20:14 | |
| *** dsneddon has joined #heat | 20:16 | |
| *** alexpilotti has quit IRC | 20:16 | |
| *** thomas_ustudio has joined #heat | 20:17 | |
| thomas_ustudio | I'm using heat to create a resource group, and I'd like each server in the group to have a unique name/hostname. Is there a way to do this? I can't find a way to get at the instance of the server in the group when passing parameters in in the template | 20:19 |
| *** HPJL_ has joined #heat | 20:20 | |
| *** HPJL has quit IRC | 20:20 | |
| zaneb | thomas_ustudio: if you don't specify a name in the server config, I think every one will get a unique name | 20:21 |
| randallburt | thomas_ustudio: there's this patch https://review.openstack.org/#/c/88636/ that I think does what you need, but I still need to fix the docs. | 20:21 |
| zaneb | though it might not be a name you like | 20:21 |
| randallburt | gotta run, be back online in a bit; sorry to chat and run. | 20:22 |
| *** SnowDust has quit IRC | 20:22 | |
| *** randallburt has quit IRC | 20:22 | |
| *** asalkeld has joined #heat | 20:22 | |
| *** JAGDISH__ has joined #heat | 20:23 | |
| *** JR___ has quit IRC | 20:24 | |
| *** e0ne has joined #heat | 20:26 | |
| *** jdandrea has quit IRC | 20:27 | |
| *** e0ne has quit IRC | 20:28 | |
| *** arunrajan has joined #heat | 20:28 | |
| *** kopparam has joined #heat | 20:28 | |
| *** arunrajan1 has joined #heat | 20:32 | |
| *** arunrajan has quit IRC | 20:32 | |
| *** arunrajan1 has quit IRC | 20:33 | |
| *** kopparam has quit IRC | 20:33 | |
| *** arunrajan has joined #heat | 20:33 | |
| *** e0ne has joined #heat | 20:36 | |
| *** e0ne has quit IRC | 20:42 | |
| *** jdandrea has joined #heat | 20:44 | |
| *** jprovazn has quit IRC | 20:45 | |
| thomas_ustudio | zaneb: I still want to be able to specify part of the name, so the role of the server is obvious | 20:48 |
| *** arbylee has quit IRC | 20:49 | |
| zaneb | the name will be derived from the name of the resource group | 20:49 |
| *** arbylee has joined #heat | 20:49 | |
| zaneb | at least, that's how it works for autoscaling groups | 20:49 |
| zaneb | I'm not 100% clear on how resource group works | 20:50 |
| thomas_ustudio | zaneb: ok, I'll give that a shot; thanks | 20:51 |
| stevebaker | morning | 20:52 |
| *** jdob has quit IRC | 20:59 | |
| *** IlyaE has quit IRC | 20:59 | |
| *** radez is now known as radez_g0n3 | 21:01 | |
| thomas_ustudio | zaneb, randallburt: Looks like it's already there: you can use %index% inside a value in a resource group (or rename the variable by setting the index_var property of the resource_group) | 21:01 |
| *** edmund1 has joined #heat | 21:02 | |
| *** edmund has quit IRC | 21:02 | |
| *** IlyaE has joined #heat | 21:03 | |
| *** SnowDust has joined #heat | 21:07 | |
| *** sarob has joined #heat | 21:15 | |
| *** sabeen has quit IRC | 21:19 | |
| *** dsneddon has quit IRC | 21:26 | |
| *** nati_uen_ has joined #heat | 21:27 | |
| *** samstav_ has joined #heat | 21:27 | |
| *** devx_ has joined #heat | 21:27 | |
| *** radez_g0` has joined #heat | 21:28 | |
| *** dsneddon has joined #heat | 21:28 | |
| *** mkollaro1 has joined #heat | 21:28 | |
| *** mkollaro has quit IRC | 21:28 | |
| *** yogeshme_ has quit IRC | 21:29 | |
| *** ccrouch1 has joined #heat | 21:29 | |
| *** Daviey_ has joined #heat | 21:30 | |
| *** erecio has quit IRC | 21:30 | |
| jdandrea | Reality check: In order to pass a list (of server uuids) into a resource group, I use a param type of comma_delimited_list and not json, correct? | 21:33 |
| *** DandyPandy has joined #heat | 21:33 | |
| *** edmund has joined #heat | 21:34 | |
| *** sarob has quit IRC | 21:34 | |
| *** devx has quit IRC | 21:34 | |
| *** radez_g0n3 has quit IRC | 21:34 | |
| *** samstav has quit IRC | 21:34 | |
| SpamapS | jdandrea: server uuids?! | 21:34 |
| SpamapS | jdandrea: you can't, AFAIK, set the UUID of a new server, can you? | 21:35 |
| SpamapS | that would be.. insane. | 21:35 |
| shardy | jdandrea: ResourceGroup takes a stack template, either inline or via referencing an existing template | 21:35 |
| shardy | https://github.com/hardys/demo_templates/blob/master/juno_summit_intro_to_heat/example3_server_with_volume_group/server_with_volume_group.yaml | 21:35 |
| jdandrea | Correction ... | 21:35 |
| *** nati_ueno has quit IRC | 21:35 | |
| *** sgordon has quit IRC | 21:35 | |
| *** blomquisg has quit IRC | 21:35 | |
| *** kevinbenton has quit IRC | 21:35 | |
| *** Daviey has quit IRC | 21:35 | |
| *** greghaynes has quit IRC | 21:35 | |
| *** DandyPandy_ has quit IRC | 21:36 | |
| *** edmund1 has quit IRC | 21:36 | |
| *** ccrouch has quit IRC | 21:36 | |
| *** Daviey_ is now known as Daviey | 21:36 | |
| *** SnowDust has quit IRC | 21:36 | |
| *** kevinbenton has joined #heat | 21:36 | |
| openstackgerrit | Jason Dunsmore proposed a change to openstack/heat: OS::Swift::Signal resource https://review.openstack.org/96947 | 21:37 |
| jdandrea | I have a provider template that, among other things, encapsulates a new resource which takes two Nova Server Resource IDs as input. In the parent, I want to pass these in (Properties from the parent become parameters in the provider template.) Is this not possible? | 21:37 |
| *** jasond` has quit IRC | 21:38 | |
| jdandrea | (I cited "resource group" by mistake - wrong terminology.) | 21:38 |
| SpamapS | jdandrea: what type are they in the provider template? Should probably be the same type. | 21:39 |
| * SpamapS puts the stress ball down | 21:39 | |
| jdandrea | SpamapS: :) | 21:39 |
| *** swygue has joined #heat | 21:40 | |
| jdandrea | SpamapS: That's a good question. I'm thinking it would be a string (e.g., VolumeAttachment takes the ID of a volume to be attached, and in the built-in resource plugin it's considered a string). | 21:41 |
| jdandrea | If I am passing more than one, however, I was thinking I'd use a comma delimited list parameter type and not the json type (overkill, perhaps). | 21:42 |
| *** sarob has joined #heat | 21:42 | |
| *** aweiteka has quit IRC | 21:42 | |
| *** mtreinish has joined #heat | 21:42 | |
| *** blamar has quit IRC | 21:43 | |
| *** blomquisg has joined #heat | 21:43 | |
| *** greghaynes has joined #heat | 21:44 | |
| *** vijendar has quit IRC | 21:45 | |
| jdandrea | So I imagine I could use properties.Schema.LIST down in the plugin, and comma_delimited_list up in the provider template parameters. | 21:45 |
| *** shuffleb1t is now known as shufflebot | 21:49 | |
| *** shufflebot has quit IRC | 21:49 | |
| *** shufflebot has joined #heat | 21:49 | |
| shardy | jdandrea: that would probably work, but you'd need to use the cfn-ish Fn::Split function to explode the comma delimited list when passing it into the plugin | 21:49 |
| shardy | it doesn't look like we have a native equivalent to that function yet AFAICS | 21:50 |
| jdandrea | Ah, ok. That was the missing link. Thank you. :) | 21:50 |
| shardy | we should fix that :) | 21:50 |
| stevebaker | on a related note, reviews welcome on list_join change https://review.openstack.org/#/c/102718/ | 21:51 |
| jdandrea | Ooh! Ahh! Thanks. | 21:52 |
| shardy | https://bugs.launchpad.net/heat/+bug/1336072 | 21:53 |
| uvirtbot | Launchpad bug 1336072 in heat "Purge remaining CFN functions in HOT" [Undecided,New] | 21:53 |
| jdandrea | So something like this in the interim: property: { "Fn::Split" : [ ",", { get_param: my_list_param } ] } | 21:53 |
| *** atiwari has joined #heat | 21:53 | |
| shardy | jdandrea, stevebaker: raised that so we can track getting native-all-the-things done, then eventually remove the cfn-isms | 21:54 |
| jdandrea | Thanks shardy. Hear hear for the -isms part. | 21:54 |
| atiwari | stevebaker, ping | 21:55 |
| stevebaker | atiwari: pong | 21:55 |
| *** arbylee has quit IRC | 21:55 | |
| atiwari | stevebaker, have some question about https://review.openstack.org/#/c/98684/. will ping you over PM | 21:56 |
| *** arbylee has joined #heat | 21:56 | |
| *** andreaf_ has quit IRC | 21:56 | |
| *** thomas_ustudio has quit IRC | 21:56 | |
| stevebaker | atiwari: no need for private ;) | 21:56 |
| atiwari | ok | 21:56 |
| stevebaker | asalkeld: are you about yet? I think this conversation is going to be about stevedore and contrib | 21:57 |
| asalkeld | hi stevebaker | 21:57 |
| atiwari | as part of https://review.openstack.org/#/c/98174/ I am planning to change the Order API | 21:57 |
| asalkeld | stevebaker, that was next on my list | 21:57 |
| asalkeld | I'll make a setup.py/cfg for each contrib | 21:57 |
| asalkeld | plugin | 21:58 |
| asalkeld | and you could then install each contib like a real package | 21:58 |
| stevebaker | asalkeld: feel free to steal this https://review.openstack.org/#/c/103441/ | 21:58 |
| atiwari | seems you are using the resource, is that OK if we break the current API resource structure .... | 21:58 |
| atiwari | ? | 21:58 |
| stevebaker | atiwari: To me, anything in contrib has zero requirement to have a stable API | 21:59 |
| asalkeld | o, cool stevebaker: great minds think alike;) | 21:59 |
| stevebaker | asalkeld: I don't think mixing stevedore and PluginManager will work well for contrib, so it will need to be converted in one hit | 21:59 |
| atiwari | stevebaker, great | 21:59 |
| atiwari | thanks | 21:59 |
| asalkeld | stevebaker, It should be fine | 21:59 |
| asalkeld | I have code to deal with the current mechanism | 22:00 |
| *** dims has quit IRC | 22:00 | |
| asalkeld | (unless you mean the client plugins mixing in with old school plugin resoruces) | 22:00 |
| *** dsneddon has quit IRC | 22:01 | |
| stevebaker | asalkeld: I'm having failing docs job which looks like a weird interaction between pbr and PluginManager http://paste.openstack.org/show/85183/ | 22:01 |
| *** jdandrea has quit IRC | 22:01 | |
| stevebaker | asalkeld: fixing the docs code to not use PluginManager and the problem should go away | 22:02 |
| *** dsneddon_ has joined #heat | 22:02 | |
| *** dsneddon_ is now known as dsneddon | 22:02 | |
| *** IlyaE has quit IRC | 22:02 | |
| asalkeld | https://review.openstack.org/#/c/103044/5/doc/source/ext/resources.py | 22:02 |
| asalkeld | stevebaker, I have that ^ | 22:03 |
| asalkeld | so it loads both | 22:03 |
| asalkeld | and produces the full list of resource docs | 22:03 |
| stevebaker | asalkeld: but as soon as the contrib/foo/setup.py exists it breaks | 22:04 |
| *** IlyaE has joined #heat | 22:04 | |
| asalkeld | stevebaker, I think it makes sense to build the docs differently when we move contirb to setup.py | 22:04 |
| asalkeld | maybe have their own docs build | 22:05 |
| stevebaker | asalkeld: should we just disable contrib docs building until we sort this all out? | 22:05 |
| asalkeld | seems ok to me | 22:05 |
| asalkeld | (short term step) | 22:05 |
| stevebaker | yeah | 22:06 |
| *** ericgoncz has quit IRC | 22:06 | |
| *** JAGDISH__ has quit IRC | 22:07 | |
| *** sarob has quit IRC | 22:10 | |
| *** sarob has joined #heat | 22:10 | |
| openstackgerrit | Steven Hardy proposed a change to openstack/heat: Use auth_token_info to initialize heat_keystoneclient sessions https://review.openstack.org/99769 | 22:13 |
| openstackgerrit | Steve Baker proposed a change to openstack/heat: Implement barbican client plugin https://review.openstack.org/98684 | 22:14 |
| openstackgerrit | Steve Baker proposed a change to openstack/heat: Use setuptools to install contrib plugins https://review.openstack.org/103441 | 22:14 |
| shardy | stevebaker: ^^ FYI just pushed that again from my branch which had the minor i8n fix in patchset2 | 22:14 |
| openstackgerrit | Steve Baker proposed a change to openstack/heat: Only support client plugins https://review.openstack.org/98688 | 22:14 |
| openstackgerrit | Steve Baker proposed a change to openstack/heat: Port rackspace clients to client plugins https://review.openstack.org/98687 | 22:14 |
| openstackgerrit | Steve Baker proposed a change to openstack/heat: Implement marconi client plugin https://review.openstack.org/98683 | 22:14 |
| *** sarob has quit IRC | 22:15 | |
| stevebaker | shardy: ok. I'm going to do more race testing on that | 22:17 |
| shardy | stevebaker: Also note https://review.openstack.org/#/c/103397/ which should have significant performance impact | 22:18 |
| shardy | That looks ready to me, but would be good to get some confidence it won't expose more races | 22:19 |
| shardy | Does sound like that combined with the auth_token patch will address the performance regression though :) | 22:19 |
| *** arbylee has quit IRC | 22:21 | |
| *** arbylee has joined #heat | 22:21 | |
| *** gondoi is now known as zz_gondoi | 22:21 | |
| stevebaker | shardy: I've approved it | 22:21 |
| *** andersonvom has quit IRC | 22:25 | |
| *** arbylee1 has joined #heat | 22:30 | |
| *** arbylee has quit IRC | 22:31 | |
| *** blamar has joined #heat | 22:32 | |
| *** shakamunyi has joined #heat | 22:33 | |
| *** sarob has joined #heat | 22:33 | |
| *** boris-42 has joined #heat | 22:36 | |
| *** bandarji has joined #heat | 22:38 | |
| *** arunrajan has quit IRC | 22:50 | |
| *** arunrajan has joined #heat | 22:51 | |
| *** arunrajan has quit IRC | 22:52 | |
| *** sarob has quit IRC | 22:54 | |
| *** arunrajan has joined #heat | 22:55 | |
| asalkeld | brb: food | 22:58 |
| *** asalkeld has quit IRC | 22:58 | |
| *** atiwari has quit IRC | 23:00 | |
| *** TravT has joined #heat | 23:08 | |
| *** sarob has joined #heat | 23:10 | |
| *** mkollaro1 has quit IRC | 23:13 | |
| *** samstav has joined #heat | 23:16 | |
| *** sabeen1 has joined #heat | 23:16 | |
| openstackgerrit | A change was merged to openstack/heat: Only do property validation in validate() https://review.openstack.org/103397 | 23:17 |
| *** samstav_ has quit IRC | 23:19 | |
| *** openstackgerrit has quit IRC | 23:19 | |
| *** openstackgerrit has joined #heat | 23:21 | |
| *** arunrajan has quit IRC | 23:25 | |
| *** arunrajan has joined #heat | 23:25 | |
| *** zhiyan_ is now known as zhiyan | 23:26 | |
| *** dims_ has joined #heat | 23:27 | |
| *** zhiyan is now known as zhiyan_ | 23:27 | |
| *** zhiyan_ is now known as zhiyan | 23:29 | |
| *** zhiyan is now known as zhiyan_ | 23:30 | |
| *** chinhuang007 has joined #heat | 23:32 | |
| *** dims_ has quit IRC | 23:32 | |
| *** david-lyle has quit IRC | 23:36 | |
| *** dims_ has joined #heat | 23:37 | |
| *** m_22 has joined #heat | 23:39 | |
| *** bandarji has quit IRC | 23:39 | |
| chinhuang007 | I want to deploy resources across multiple regions. Do I need multiple templates or can I have a heat template spanning multiple regions? | 23:41 |
| *** m_22 has left #heat | 23:41 | |
| *** DandyPandy has quit IRC | 23:48 | |
| *** shakamunyi has quit IRC | 23:52 | |
| *** kebray has quit IRC | 23:53 | |
| *** shakamunyi has joined #heat | 23:53 | |
| *** shakamunyi has quit IRC | 23:53 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!