*** Qiming_ has quit IRC | 00:14 | |
*** Qiming has joined #senlin | 00:25 | |
*** Qiming has quit IRC | 00:30 | |
*** pratikmallya has joined #senlin | 00:51 | |
*** pratikmallya has quit IRC | 00:56 | |
*** pratikmallya has joined #senlin | 01:08 | |
*** xujun has quit IRC | 01:16 | |
*** zhenguo has joined #senlin | 01:22 | |
*** Yanyanhu has joined #senlin | 01:31 | |
openstackgerrit | Zhenguo Niu proposed openstack/senlin-dashboard: Remove "builtins = _" from tox.ini https://review.openstack.org/261133 | 02:03 |
---|---|---|
*** elynn has joined #senlin | 02:04 | |
elynn | morning | 02:04 |
*** yuanying has quit IRC | 02:08 | |
*** yuanying has joined #senlin | 02:20 | |
*** yuanying_ has joined #senlin | 02:24 | |
*** yuanying has quit IRC | 02:26 | |
*** yuanying has joined #senlin | 02:26 | |
*** yuanyin__ has joined #senlin | 02:28 | |
*** yuanying_ has quit IRC | 02:29 | |
*** yuanying has quit IRC | 02:31 | |
*** yuanyin__ has quit IRC | 02:39 | |
openstackgerrit | Ethan Lynn proposed openstack/senlin: Set status to updating when execute resize action https://review.openstack.org/261138 | 02:44 |
*** yuanying has joined #senlin | 02:46 | |
*** pratikmallya has quit IRC | 02:51 | |
*** pratikmallya has joined #senlin | 03:23 | |
*** yuanying has quit IRC | 03:25 | |
elynn | Hi Yanyanhu | 03:39 |
Yanyanhu | hello | 03:39 |
elynn | https://bugs.launchpad.net/senlin/+bug/1528996 | 03:40 |
openstack | Launchpad bug 1528996 in senlin "Can't delete heat node" [Undecided,New] - Assigned to Ethan Lynn (ethanlynn) | 03:40 |
*** yuanying has joined #senlin | 03:40 | |
Yanyanhu | this is weird | 03:40 |
elynn | I'm not sure it should be solve at sdk or senlin side... | 03:40 |
elynn | Yes. | 03:40 |
Yanyanhu | let me check the code, I guess show_deleted should not be set by default? | 03:41 |
elynn | I think it's sdk's bug... | 03:41 |
Yanyanhu | yes, guess so | 03:42 |
Yanyanhu | according to the comment of find_stack function call, None should be returned if stack is not found and ignore_missing=True | 03:43 |
elynn | http://paste.openstack.org/show/482638/ it's the log of senlin-engine | 03:43 |
elynn | not found response contains a redirect link and sdk follows it and get the stack... | 03:44 |
Yanyanhu | oh | 03:44 |
Yanyanhu | sdk or keystoneauth's problem I think | 03:45 |
Yanyanhu | if so, similar error should happen when sdk talking with other services? | 03:46 |
elynn | No | 03:46 |
elynn | I test nova | 03:46 |
elynn | it works fine... | 03:46 |
elynn | just don't know why | 03:47 |
Yanyanhu | hmm, a little weird | 03:48 |
elynn | I think that makes functional tests fail recently. | 03:49 |
Yanyanhu | yes, I noticed functional test failed | 03:50 |
elynn | Will have lunch now, will be back later and dig more | 03:50 |
Yanyanhu | so you mean the redirection causes this issue? | 03:50 |
Yanyanhu | ok, will also look at it | 03:50 |
openstackgerrit | Merged openstack/senlin: install.sh: stop using deprecated option group for rabbit https://review.openstack.org/260937 | 04:00 |
*** yuanying_ has joined #senlin | 04:05 | |
*** yuanying_ has quit IRC | 04:07 | |
*** yuanying_ has joined #senlin | 04:07 | |
*** yuanying has quit IRC | 04:07 | |
*** yuanying has joined #senlin | 04:09 | |
*** yuanying_ has quit IRC | 04:12 | |
*** yuanying_ has joined #senlin | 04:22 | |
*** yuanying has quit IRC | 04:23 | |
*** yuanying has joined #senlin | 04:25 | |
*** yuanying_ has quit IRC | 04:25 | |
*** yuanying_ has joined #senlin | 04:27 | |
*** yuanying_ has quit IRC | 04:28 | |
*** yuanying_ has joined #senlin | 04:28 | |
*** yuanying has quit IRC | 04:28 | |
*** yuanying has joined #senlin | 04:38 | |
*** yuanying_ has quit IRC | 04:40 | |
*** yuanying has quit IRC | 04:43 | |
Yanyanhu | hi, elynn, are you around? | 04:50 |
elynn | Hi Yanyanhu | 04:50 |
Yanyanhu | about bug 1528996 | 04:50 |
openstack | bug 1528996 in senlin "Can't delete heat node" [Undecided,Confirmed] https://launchpad.net/bugs/1528996 - Assigned to Ethan Lynn (ethanlynn) | 04:50 |
Yanyanhu | could it be heat's bug? | 04:51 |
Yanyanhu | http://9.186.107.212:8004/v1/1d567ed4ef51453a85545f018b68c26d/stacks/n-01-Gy1hbCr3/ce374e79-a10b-4258-88f7-3784b06ba5a5 | 04:51 |
*** yuanying has joined #senlin | 04:52 | |
Yanyanhu | I guess GET request to the above link should not get stack detail returned? | 04:52 |
elynn | I test with heatclient, the response body doesn't contain redirect link | 04:52 |
Yanyanhu | after the redirection happened? | 04:52 |
Yanyanhu | oh | 04:52 |
Yanyanhu | weird | 04:52 |
elynn | But heatclient request this link http://9.186.107.212:8004/v1/1d567ed4ef51453a85545f018b68c26d/stacks/n-01-Gy1hbCr3/ | 04:53 |
elynn | without that id | 04:53 |
Yanyanhu | maybe that's is because you request for a legal url? | 04:53 |
Yanyanhu | yes, this is a legal url which supported by Heat | 04:53 |
Yanyanhu | http://developer.openstack.org/api-ref-orchestration-v1.html#stack_find | 04:53 |
Yanyanhu | but this one is not? curl -g -i -X GET http://9.186.107.212:8004/v1/1d567ed4ef51453a85545f018b68c26d/stacks/ce374e79-a10b-4258-88f7-3784b06ba5a5 | 04:54 |
Yanyanhu | in which stack id is located after 'stacks/' | 04:54 |
elynn | hmm http://developer.openstack.org/api-ref-orchestration-v1.html#stack_show | 04:55 |
elynn | it's for stack_show | 04:55 |
Yanyanhu | so I guess there could be two problems here: 1)sdk should not request such url in which stack id is located after str 'stacks/' when finding stack; | 04:56 |
Yanyanhu | 2) why heat return redirect info in header after receive such a request... | 04:56 |
Yanyanhu | since if heat treats the id string as name, it should return 404 rather than 302 I think | 04:57 |
elynn | From the log | 04:59 |
elynn | I don't know sdk request what url, I just see it try to request redirect link. | 04:59 |
elynn | let me have a try... | 05:00 |
Yanyanhu | yes, I guess at the first time, sdk requests this kind of url 'http://9.186.107.212:8004/v1/1d567ed4ef51453a85545f018b68c26d/stacks/ce374e79-a10b-4258-88f7-3784b06ba5a5' | 05:00 |
Yanyanhu | let me paste some of my engine log | 05:00 |
*** xuhaiwei_ has quit IRC | 05:01 | |
*** yuanying has quit IRC | 05:01 | |
Yanyanhu | http://paste.openstack.org/show/482639/ | 05:01 |
Yanyanhu | I guess line1 has the original url sdk requested? | 05:01 |
Yanyanhu | actually, keystoneauth1 requested | 05:02 |
openstackgerrit | Zhenguo Niu proposed openstack/senlin-dashboard: url template tag change https://review.openstack.org/261152 | 05:02 |
elynn | Hi Yanyanhu | 05:04 |
*** yuanying has joined #senlin | 05:05 | |
elynn | I use restful tool to request this link and I can get the stack details... | 05:05 |
Yanyanhu | ... | 05:05 |
Yanyanhu | but we shouldn't I think ? | 05:05 |
elynn | Heat treat it as stack_show | 05:06 |
Yanyanhu | can you paste the url you requested | 05:06 |
elynn | not stack_find | 05:06 |
*** yuanying has quit IRC | 05:06 | |
elynn | http://9.123.137.235:8004/v1/1e6a5c5d14f64768991f691616d3ac67/stacks/35d2103b-d831-449e-b380-5d5cc313c249 | 05:06 |
Yanyanhu | but I think the url of stack_show should be in the following format: /v1/{tenant_id}/stacks/{stack_name}/{stack_id} | 05:07 |
Yanyanhu | without stack_name segment, it works as well? | 05:07 |
elynn | Heat send back a redirect link to correct it. | 05:07 |
Yanyanhu | then the api doc of heat is wrong... | 05:08 |
Yanyanhu | http://developer.openstack.org/api-ref-orchestration-v1.html#stack_find | 05:08 |
elynn | And my restful tool follows it I think... | 05:08 |
Yanyanhu | according to the description, only stack_name is accepted... | 05:08 |
Yanyanhu | Request parameters | 05:08 |
Yanyanhu | Parameter Style Type Description | 05:08 |
Yanyanhu | tenant_id URI String | 05:08 |
Yanyanhu | The ID of the tenant. A tenant is also known as an account or project. | 05:08 |
Yanyanhu | stack_name URI String | 05:08 |
Yanyanhu | The name of a stack. | 05:08 |
elynn | If I request http://9.123.137.235:8004/v1/1e6a5c5d14f64768991f691616d3ac67/stacks/stack_name | 05:08 |
elynn | I can get en error. | 05:08 |
Yanyanhu | .. | 05:08 |
Yanyanhu | sigh... | 05:08 |
*** yuanying has joined #senlin | 05:09 | |
Yanyanhu | so the heat API-ref doc need fix I think... | 05:10 |
elynn | not sure if heat cores willing to change this behavior... | 05:10 |
Yanyanhu | and also stack resource implementation in sdk | 05:10 |
Yanyanhu | maybe just fix the documentation? | 05:10 |
elynn | They need to consider backward compability... | 05:11 |
elynn | Yanyanhu, What document? | 05:11 |
Yanyanhu | elynn, yes, just the document and the actual implementation is mismatched... | 05:11 |
Yanyanhu | http://developer.openstack.org/api-ref-orchestration-v1.html#stack_find | 05:11 |
Yanyanhu | this one | 05:11 |
Yanyanhu | and I found show_delete is True by default when getting stack by ID | 05:14 |
Yanyanhu | and there is no option to change this default behavior when requesting stack get API... | 05:15 |
openstackgerrit | Zhenguo Niu proposed openstack/senlin-dashboard: Add Django1.8 to tests https://review.openstack.org/261154 | 05:15 |
Yanyanhu | so either change related implementation in heat API (I guess this is not easy), or revise heat stack resource type in sdk | 05:16 |
Yanyanhu | to define special logic for find resource call of heat stack | 05:17 |
*** xuhaiwei has joined #senlin | 05:20 | |
elynn | Yanyanhu, I don't find the doc error... | 05:27 |
Yanyanhu | ah... let me check it again | 05:28 |
elynn | Yanyanhu, I would like to suggest to change sdk... | 05:28 |
Yanyanhu | yes, this is also what I prefer | 05:28 |
Yanyanhu | http://developer.openstack.org/api-ref-orchestration-v1.html#stack_find | 05:28 |
Yanyanhu | this doc link, and description of the second requests parameter | 05:29 |
elynn | I thinks it's correct, don't know which line you think is not correct. | 05:29 |
Yanyanhu | I mean actually heat stack find API accept stack_id here | 05:29 |
elynn | oh, I got your point. | 05:30 |
Yanyanhu | not stack_name | 05:30 |
Yanyanhu | yes, so I guess they mean to implement the API like this | 05:30 |
Yanyanhu | just doc is inconsistent with the actual implementation | 05:30 |
elynn | Let me try with exist stack... | 05:31 |
Yanyanhu | ok | 05:31 |
elynn | heat treat name and id as different... sigh | 05:35 |
elynn | You can not find stack by name if deleted, but you can find it by id | 05:36 |
elynn | And you can not show a deleted stack by name if it's deleted. | 05:40 |
elynn | But you can show it by id.... | 05:40 |
Yanyanhu | ... | 05:40 |
Yanyanhu | yes | 05:40 |
elynn | I just remember somebody raise this problem before, but cores ignore it | 05:41 |
Yanyanhu | so if you try to 'find' a stack in heat, you have to list stacks and filt with given name... | 05:41 |
Yanyanhu | using stack_id won't work since no way to change the default behavior(show_deleted=True) of stack get API | 05:42 |
Yanyanhu | actually, IMHO, heat should expose show_deleted option for stack get API :) | 05:42 |
Yanyanhu | it will be very helpful | 05:43 |
elynn | Yanyanhu, I agree with you. | 05:43 |
Yanyanhu | anyway, before heat can change this design, we may need to fix this issue by proposing patch to sdk | 05:44 |
elynn | Yanyanhu, yes, I also report a bug at heat side. https://bugs.launchpad.net/heat/+bug/1529001 | 05:51 |
openstack | Launchpad bug 1529001 in heat "find_stack api treat stack_name and id differently" [Undecided,New] | 05:51 |
Yanyanhu | thanks :) | 05:51 |
elynn | Do you know how to fix it in sdk? | 05:51 |
Yanyanhu | sorry, elynn, just left for a while | 06:03 |
*** yuanying has quit IRC | 06:03 | |
Yanyanhu | elynn, I think we may need to overwrite the find resource call of heat stack resource | 06:03 |
Yanyanhu | currently, the find method defined in parent Resource class is used | 06:04 |
Yanyanhu | let me find the code | 06:04 |
Yanyanhu | http://git.openstack.org/cgit/openstack/python-openstacksdk/tree/openstack/resource.py#n881 | 06:04 |
Yanyanhu | this is the Resource class | 06:05 |
Yanyanhu | may need to add a overwriting one to here http://git.openstack.org/cgit/openstack/python-openstacksdk/tree/openstack/orchestration/v1/stack.py | 06:05 |
*** yuanying has joined #senlin | 06:06 | |
Yanyanhu | since most of heat stack support in sdk was implemented by qiming, you can also get some advise from him I think | 06:08 |
Yanyanhu | but I feel overwriting 'find' resource call could be the easist way to fix this issue before heat API can be revised... | 06:09 |
elynn | Maybe overwrite get_by_id is ok. | 06:15 |
Yanyanhu | elynn, yes | 06:28 |
Yanyanhu | hi, will leave for a meeting, ttyl | 06:28 |
*** pratikmallya has quit IRC | 07:38 | |
elynn | Submit a patch to openstacksdk, but don't know how to write a unittests... | 08:33 |
openstackgerrit | Ethan Lynn proposed openstack/senlin: [Do not review] Test heat driver https://review.openstack.org/261223 | 08:41 |
*** elynn has quit IRC | 09:05 | |
*** Yanyanhu has quit IRC | 09:41 | |
*** yuanying has quit IRC | 10:41 | |
openstackgerrit | Merged openstack/senlin-dashboard: Remove "builtins = _" from tox.ini https://review.openstack.org/261133 | 14:43 |
openstackgerrit | Merged openstack/senlin-dashboard: url template tag change https://review.openstack.org/261152 | 14:47 |
openstackgerrit | Merged openstack/senlin-dashboard: Add Django1.8 to tests https://review.openstack.org/261154 | 14:48 |
openstackgerrit | Zhenguo Niu proposed openstack/senlin-dashboard: Remove django<1.8 from tests https://review.openstack.org/261341 | 15:14 |
openstackgerrit | Merged openstack/senlin-dashboard: Remove django<1.8 from tests https://review.openstack.org/261341 | 15:28 |
*** zhenguo has quit IRC | 15:51 | |
*** bdrich_ has joined #senlin | 16:51 | |
*** pratikmallya has joined #senlin | 17:47 | |
*** bdrich_ has quit IRC | 17:48 | |
*** openstackstatus has quit IRC | 18:05 | |
*** openstack has joined #senlin | 18:09 | |
*** pratikmallya has quit IRC | 19:14 | |
*** pratikmallya has joined #senlin | 20:40 | |
*** pratikmallya has quit IRC | 22:17 | |
*** pratikmallya has joined #senlin | 22:21 | |
*** yuanying has joined #senlin | 23:05 | |
*** pratikmallya has quit IRC | 23:40 | |
*** pratikmallya has joined #senlin | 23:44 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!