| *** lvdongbing has joined #senlin | 00:30 | |
| *** zhenguo has joined #senlin | 01:20 | |
| *** elynn has joined #senlin | 01:55 | |
| *** Yanyanhu has joined #senlin | 01:57 | |
| *** xuhaiwei has joined #senlin | 01:59 | |
| openstackgerrit | Merged openstack/senlin: Fix the log message issue in engine.service.profile_create https://review.openstack.org/239619 | 02:08 |
|---|---|---|
| *** yangyapeng has joined #senlin | 02:08 | |
| openstackgerrit | Merged openstack/senlin: Use assertTrue/False instead of assertEqual(T/F) https://review.openstack.org/239609 | 02:11 |
| *** elynn has quit IRC | 02:13 | |
| openstackgerrit | Merged openstack/senlin: Use assertIsNone(observed) instead of assertEqual(None, observed) https://review.openstack.org/239598 | 02:17 |
| *** elynn has joined #senlin | 02:20 | |
| openstackgerrit | Merged openstack/senlin: Add oslo_config "IPOpt" and "PortOpt" support https://review.openstack.org/239350 | 02:21 |
| *** Qiming has joined #senlin | 02:21 | |
| *** Qiming has quit IRC | 02:21 | |
| *** Qiming has joined #senlin | 02:22 | |
| Qiming | morning | 02:22 |
| Yanyanhu | morning | 02:22 |
| Yanyanhu | hi, Qiming, I think maybe we should updated TODO.rst based on our discussion result in meetup?https://etherpad.openstack.org/p/senlin-mitaka-meetup | 02:23 |
| Qiming | yes | 02:23 |
| Yanyanhu | ok, will try to look through all those workitems and fill them into TODO files | 02:23 |
| Qiming | cool, thx | 02:24 |
| Yanyanhu | np :) | 02:24 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Add config option for name unique https://review.openstack.org/234051 | 02:27 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Add config option for name unique https://review.openstack.org/234051 | 02:28 |
| openstackgerrit | lvdongbing proposed openstack/senlin: API resource names should not include underscores https://review.openstack.org/240740 | 02:29 |
| *** TianYuan has joined #senlin | 02:33 | |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Add config option for name uniqueness https://review.openstack.org/234051 | 02:45 |
| *** TianYuan has quit IRC | 02:53 | |
| *** TianYuan has joined #senlin | 02:54 | |
| *** elynn_ has joined #senlin | 03:05 | |
| *** elynn has quit IRC | 03:08 | |
| elynn_ | morning | 03:08 |
| openstackgerrit | lvdongbing proposed openstack/senlin: API resource names should not include underscores https://review.openstack.org/240740 | 03:27 |
| *** Qiming has quit IRC | 03:32 | |
| openstackgerrit | Merged openstack/python-senlinclient: Update gitignore file https://review.openstack.org/239593 | 04:37 |
| *** ldb_ has joined #senlin | 04:47 | |
| *** lvdongbing has quit IRC | 04:49 | |
| *** Qiming has joined #senlin | 05:01 | |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Fix setup-service script for RC/role checking https://review.openstack.org/240760 | 05:24 |
| *** yangyapeng has quit IRC | 05:31 | |
| *** TianYuan has quit IRC | 05:35 | |
| *** yangyapeng has joined #senlin | 05:51 | |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Update TODO item list https://review.openstack.org/238432 | 06:06 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Update TODO item list https://review.openstack.org/240764 | 06:06 |
| openstackgerrit | Merged openstack/senlin: Update TODO item list https://review.openstack.org/238432 | 07:28 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Update TODO item list https://review.openstack.org/240764 | 07:35 |
| *** lawrancejing has joined #senlin | 07:48 | |
| *** lawrancejing has quit IRC | 07:50 | |
| openstackgerrit | Liuqing Jing proposed openstack/senlin-dashboard: Add python-senlinclient in requirements.txt https://review.openstack.org/227249 | 08:06 |
| Qiming | elynn_, there? | 08:07 |
| elynn_ | Hi | 08:08 |
| Qiming | https://review.openstack.org/#/c/240776 | 08:08 |
| elynn_ | yes? | 08:08 |
| Qiming | about this one, I think we should put it on hold | 08:08 |
| Qiming | from the code you can see the logic is not right | 08:08 |
| elynn_ | ok...but why | 08:08 |
| Qiming | it is senlinclient's problem | 08:09 |
| Qiming | or we can say it is sdk's problem | 08:09 |
| elynn_ | You mean those cluster_* functions? | 08:09 |
| Qiming | yes | 08:09 |
| elynn_ | Yes, I notices that... | 08:09 |
| Qiming | from a service/software that invokes senlinclient | 08:09 |
| Qiming | the api should reads straightforward | 08:10 |
| Qiming | for example: | 08:10 |
| Qiming | clusters = self.client().clusters() | 08:10 |
| elynn_ | Yes, that would be elegent. | 08:11 |
| Qiming | cluster = self.client().create_cluster(params) | 08:11 |
| elynn_ | So I should change senlinclient first, isn't it? | 08:11 |
| Qiming | this has been a mistake for historical reasons | 08:11 |
| Qiming | we were misled by the sdk examples | 08:11 |
| Qiming | but anyway, we should stick to the suggested way to invoke sdk, including the senlin service | 08:12 |
| Qiming | I'm thinking that we should put this work on a high priority | 08:12 |
| Qiming | because the senlin-dashboard has been suffering from the same problem | 08:12 |
| elynn_ | But I notice that openstacksdk doesn't support *profile* functions yet, will it block? | 08:13 |
| Qiming | yes, I'll push sdk guys for a review | 08:13 |
| elynn_ | Yes, I copy codes from senlin-dashboard... | 08:13 |
| Qiming | https://review.openstack.org/#/c/209743/ | 08:14 |
| Qiming | it is there but not yet reviewed | 08:14 |
| Qiming | so, imho, we have two things to do | 08:15 |
| Qiming | one is to refine senlinclient to expose a much easier interface for other service to invoke | 08:15 |
| Qiming | this can be done even before the sdk has adopted all senline abstractions(resources) | 08:15 |
| Qiming | the other one is we push all senlin abstractions/resources to sdk, so we can throw away a huge amount of code from senlinclient | 08:16 |
| Qiming | in fact, this is one of the reasons why I haven't been put much efforts into the senlinclient test cases | 08:16 |
| elynn_ | ok, I will try to redefine senlinclient first. | 08:17 |
| Qiming | yep, that would be great | 08:17 |
| Qiming | in the long run, senlinclient will disappear, hopefully | 08:18 |
| Qiming | for command line, people should use openstackclient | 08:18 |
| Qiming | for api invocation, people should use openstacksdk | 08:18 |
| Qiming | but .... we are not living in a perfect world | 08:18 |
| Qiming | still a lot dirty things to do before we are getting there | 08:18 |
| elynn_ | yes, hope I can help you with it ;) | 08:19 |
| Qiming | sure | 08:19 |
| Qiming | help the team, :) | 08:19 |
| Qiming | it really doesn't make senses for each openstack project to have its own xxxclient project | 08:19 |
| elynn_ | Another question is that is it ok to add senlinclient into requirements.txt in heat? | 08:20 |
| elynn_ | Since senlin is not official yet, it's still stackforge even if the namespace changed. | 08:20 |
| Qiming | yes, it is already in global requirements | 08:21 |
| Qiming | it doesn't matter whether senlin is an official project, senlinclient is a valid dependency because it is in global requirements | 08:21 |
| Qiming | elynn_, how about this | 08:22 |
| elynn_ | ok, got it. | 08:22 |
| elynn_ | ? | 08:22 |
| Qiming | I will start a patch doing revisions to senlinclient | 08:23 |
| Qiming | since we have a lot of commands to rewrite, will need helps from you (and others) to get this change done asap | 08:23 |
| Qiming | I'll try illustrate how the client should be refactored with one or two examples | 08:24 |
| elynn_ | Cool, sounds good to me. | 08:25 |
| Qiming | then we can sync on the particular commands each of us can work on | 08:25 |
| elynn_ | yes, but how will you illustrate? | 08:25 |
| elynn_ | On text or f2f? | 08:25 |
| Qiming | illustrate using code | 08:35 |
| openstackgerrit | Merged openstack/senlin-dashboard: Change repository from stackforge to openstack https://review.openstack.org/236686 | 08:37 |
| elynn_ | ok | 08:38 |
| openstackgerrit | Merged openstack/senlin: Fix setup-service script for RC/role checking https://review.openstack.org/240760 | 08:46 |
| Qiming | elynn_, https://review.openstack.org/#/c/194496/4/requirements.txt | 09:10 |
| Qiming | typo | 09:10 |
| elynn_ | oops... | 09:10 |
| elynn_ | I should directly copy from global requirements. | 09:11 |
| Qiming | :) | 09:11 |
| *** ldb_ has quit IRC | 09:42 | |
| openstackgerrit | Qiming Teng proposed openstack/python-senlinclient: Revised client interface https://review.openstack.org/240805 | 09:43 |
| Qiming | elynn_, ^ | 09:43 |
| Qiming | this is what I meant to rework the client side interface | 09:44 |
| elynn_ | you are so quick. | 09:44 |
| Qiming | just some starting points | 09:45 |
| elynn_ | Qiming, If do cluster_create, do we need to add related functions to senlinclient? | 09:51 |
| elynn_ | Like what you did to profile operation. | 09:51 |
| Qiming | yes | 09:51 |
| Qiming | that will make your heat side code stable | 09:51 |
| Qiming | we will have some mess to clean up and some negotiations with SDK, but I don't want heat side code to be changed this way and that way | 09:52 |
| elynn_ | If openstacksdk is stable in future, do we need to clean all of these new added codes ? | 09:53 |
| Yanyanhu | hi, Qiming, just looked this patch, https://review.openstack.org/#/c/240805/1/senlinclient/v1/client.py what you mean by the comment of line 39 and 40? | 09:54 |
| Yanyanhu | you mean after all senlin resources are adopted by openstacksdk, other software(e.g. heat) should directly invoke openstacksdk to talk with senlin service? | 09:54 |
| Qiming | yes, ideally | 09:55 |
| Qiming | they can choose not to | 09:55 |
| Qiming | :) | 09:55 |
| Yanyanhu | understand | 09:55 |
| Qiming | that is one thing for senlin users | 09:55 |
| Qiming | inside senlinclient, we will need to revise the translation layer again | 09:55 |
| Qiming | we talk directly to sdk, instead of maintaining our own models | 09:56 |
| Yanyanhu | yes, that's why you add the comments in line 82-86 | 09:57 |
| Qiming | yes | 09:57 |
| *** zhenguo has quit IRC | 09:57 | |
| Yanyanhu | BTW, just as you said, I also think the methods defined in line 43-81 maybe helpful to keep the interface stable | 09:57 |
| Qiming | they are only useful before we are not supported by sdk directly | 09:57 |
| Qiming | right | 09:58 |
| Yanyanhu | since currently, the way invoking interfaces of senlinclient is a little different from the way using other xxxclients | 09:58 |
| Qiming | I was writing those logics to be very similar to the interface you will expect from openstacksdk directly | 09:58 |
| Yanyanhu | ok | 09:59 |
| elynn_ | Ok. I got your point. | 09:59 |
| Qiming | the most weird thing is client needs to know 'models.xxx', that is stupid | 09:59 |
| Yanyanhu | yes | 09:59 |
| Qiming | it was acceptable if only shell calls the client | 09:59 |
| Qiming | it is definitely not okay for other software, i.e. Heat, senlin-dashboard ... | 09:59 |
| Yanyanhu | yes | 10:00 |
| Qiming | senlin's proposal to big tent will be voted this week | 10:00 |
| Qiming | just got email from tttttx | 10:00 |
| Qiming | got to be there early in the morning, 4am, sigh | 10:00 |
| Yanyanhu | so we just put this layer in client.py(maybe temporarily) and let user to choose which way to use(invoking those interfaces defined in this layer or invoking sdk directly) | 10:00 |
| Yanyanhu | ... | 10:01 |
| Qiming | they will invoke senlinclient I think | 10:01 |
| Yanyanhu | hmm, also think so | 10:01 |
| Qiming | but I won't suggest people to do that in the long run | 10:01 |
| Yanyanhu | at least in this stage | 10:01 |
| Yanyanhu | yep | 10:01 |
| Qiming | the problem is serious in Heat | 10:01 |
| Qiming | they should not invoke all those other clients | 10:02 |
| Qiming | it ... is messy | 10:02 |
| Qiming | hard to keep track on all the clients an orhestrator needs to follow | 10:02 |
| Yanyanhu | yes, that's the meaning of openstackclient and openstacksdk :) | 10:02 |
| Qiming | sadly, not so many people today really cares about sdk | 10:02 |
| elynn_ | May be because sdk is not in big tent? | 10:04 |
| Qiming | I should have gone to a temple back in Tokyo | 10:04 |
| Yanyanhu | that's possible | 10:04 |
| Qiming | to pray | 10:04 |
| Qiming | sdk ... Brian has his own thoughts | 10:04 |
| Qiming | he doesn't believe SDK should be in the big tent | 10:04 |
| Qiming | I don't quite understand the logic behind | 10:04 |
| Yanyanhu | ...sigh | 10:05 |
| Qiming | it was about some rackspace, mirantis concerns .... whatever | 10:05 |
| Qiming | from a community job's standing point, I'd really like to see that happening | 10:06 |
| Yanyanhu | yes, but most companies usually don't think like this | 10:07 |
| Qiming | right, they have their own agenda maybe | 10:08 |
| Qiming | hungry now | 10:09 |
| Yanyanhu | :) | 10:09 |
| Yanyanhu | I'm not :P | 10:09 |
| Yanyanhu | although I don't know why | 10:09 |
| Qiming | have to go home before starved to death | 10:09 |
| Qiming | your container is small | 10:10 |
| Qiming | :) | 10:10 |
| Yanyanhu | guess so, haha | 10:10 |
| Qiming | ttyl | 10:10 |
| Yanyanhu | see U tomorrow | 10:10 |
| elynn_ | See U~ | 10:11 |
| *** elynn_ has quit IRC | 10:12 | |
| *** Qiming has quit IRC | 10:14 | |
| *** yangyapeng has quit IRC | 10:49 | |
| *** Yanyanhu has quit IRC | 10:52 | |
| *** Qiming has joined #senlin | 11:08 | |
| openstackgerrit | Liuqing Jing proposed openstack/senlin-dashboard: Add python-senlinclient in requirements.txt https://review.openstack.org/227249 | 11:22 |
| *** Liuqing has joined #senlin | 11:44 | |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Move policy detach from action to cluster https://review.openstack.org/240685 | 11:56 |
| *** Qiming has quit IRC | 12:22 | |
| *** Qiming has joined #senlin | 12:22 | |
| *** Liuqing_ has joined #senlin | 13:27 | |
| *** Liuqing has quit IRC | 13:30 | |
| openstackgerrit | Merged openstack/senlin-dashboard: Add python-senlinclient in requirements.txt https://review.openstack.org/227249 | 15:05 |
| *** Liuqing_ has quit IRC | 15:10 | |
| *** pratikma_ has joined #senlin | 15:33 | |
| pratikma_ | hello! How is senlin different from using resource groups in openstack heat? | 15:33 |
| Qiming | hi, pratikma_ | 15:33 |
| Qiming | senlin is designed to provide more knobs for user to control a group of things | 15:34 |
| Qiming | with heat, you will have to go 'stack-update' to change anything inside a resource-group | 15:35 |
| Qiming | with senlin, you will get more APIs on doing membership management, policy management, etc. | 15:35 |
| pratikma_ | Qiming: thanks! So, I'm guessing that one would use RG for simple cases, and when more management is required, then use Senlin? | 15:36 |
| Qiming | pratikma_, yes | 15:36 |
| Qiming | we are working on adding senlin 'cluster' resource type support into heat | 15:37 |
| Qiming | but due to heat's design, you won't be able to 'enjoy' all flexibilities you get from senlin, :) | 15:37 |
| pratikma_ | Qiming: I guess I'm having trouble understanding that if Senlin can manage heat stacks, then you're also proposing heat can manage Senlin clusters? | 15:39 |
| Qiming | that is a question a lot of people asked | 15:40 |
| pratikma_ | Qiming: btw it was the patch for Senlin resources that piqued my curiosity about Senlin :) | 15:40 |
| Qiming | if you have ever written some C code | 15:40 |
| Qiming | you can think of a heat stack as a 'struct', senlin cluster as an array | 15:40 |
| Qiming | you can have arrays in your struct, you can also build arrays of struct | 15:41 |
| pratikma_ | ah, I see | 15:44 |
| *** lkarm has joined #senlin | 15:44 | |
| Qiming | night, all | 16:02 |
| *** Qiming has quit IRC | 16:02 | |
| *** pratikma_ has quit IRC | 17:39 | |
| *** pratikmallya has joined #senlin | 18:05 | |
| *** pratikmallya has quit IRC | 18:06 | |
| *** pratikmallya has joined #senlin | 18:06 | |
| *** pratikma_ has joined #senlin | 18:33 | |
| *** pratikmallya has quit IRC | 18:36 | |
| *** lkarm has quit IRC | 19:55 | |
| *** lkarm has joined #senlin | 19:55 | |
| *** lkarm has quit IRC | 19:56 | |
| *** lkarm has joined #senlin | 19:57 | |
| *** pratikma_ has quit IRC | 20:45 | |
| *** pratikmallya has joined #senlin | 21:08 | |
| *** lkarm has quit IRC | 21:18 | |
| *** jdandrea has joined #senlin | 21:22 | |
| *** lkarm has joined #senlin | 21:23 | |
| *** pratikma_ has joined #senlin | 21:24 | |
| *** pratikmallya has quit IRC | 21:27 | |
| *** pratikma_ has quit IRC | 21:53 | |
| *** pratikmallya has joined #senlin | 22:31 | |
| *** lkarm has quit IRC | 22:32 | |
| *** lkarm has joined #senlin | 22:33 | |
| *** lkarm has quit IRC | 22:37 | |
| *** pratikma_ has joined #senlin | 22:52 | |
| *** pratikmallya has quit IRC | 22:54 | |
| *** lkarm has joined #senlin | 23:18 | |
| *** lkarm has quit IRC | 23:23 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!