Tuesday, 2015-12-01

*** Qiming has quit IRC00:11
*** Qiming has joined #senlin00:11
*** rebase has quit IRC00:15
*** Qiming_ has joined #senlin00:20
*** Qiming has quit IRC00:20
*** openstackstatus has joined #senlin00:26
*** ChanServ sets mode: +v openstackstatus00:26
*** Qiming_ has quit IRC00:27
*** Qiming_ has joined #senlin00:27
*** Qiming__ has joined #senlin00:28
*** Qiming__ is now known as Qiming00:31
Qimingwelcome, openstackstatus00:31
*** Qiming_ has quit IRC00:32
*** rebase has joined #senlin00:34
*** rebase has quit IRC00:51
*** xuhaiwei has joined #senlin00:53
*** zhenguo has joined #senlin01:19
*** openstackgerrit has quit IRC01:22
*** openstackgerrit has joined #senlin01:22
*** Yanyanhu has joined #senlin01:27
openstackgerritMerged openstack/senlin: Minor tweaks to policy-update log messages  https://review.openstack.org/25111301:43
openstackgerritMerged openstack/senlin: Fix action request method and URI  https://review.openstack.org/25121201:43
*** lixinhui has joined #senlin02:08
*** Yanyanhu has quit IRC02:17
*** elynn has joined #senlin02:19
elynnmorning ~02:41
elynnI can't connect to freenode yesterday and I don't know why...02:48
xuhaiweime too02:48
xuhaiweifrom yesterday afternoon about 15:0002:49
elynnAh, Seems freenode is not stable yesterday.02:49
xuhaiweibut my colleague can access to it02:49
xuhaiweiI am using browse not irc tools02:50
elynnI cant login freenode via neither web not xchat...02:50
elynnAlways get an connection failure02:51
xuhaiweiyes, the server is strange yesterday02:54
*** pm90_ has quit IRC02:55
Qimingfreenode was under DDOS attack yesterday03:04
openstackgerritQiming Teng proposed openstack/senlin: Fix documentation for conformance  https://review.openstack.org/25123003:08
openstackgerritMerged openstack/senlin: Decide auth_plugin type in create_connection  https://review.openstack.org/25126203:08
elynnOpenStack should create a project for anti-DDOS  :D03:11
openstackgerritMerged openstack/python-senlinclient: Fix action api calls  https://review.openstack.org/25121903:12
Qimingthat would be a sub-project of neutron?03:12
elynnen, your idea sounds more reasonable03:14
elynnIs there any NFV project to do that?03:14
*** yuanying has quit IRC03:23
openstackgerritYanyan Hu proposed openstack/senlin: Add more configurable parameters to wait_for_server  https://review.openstack.org/25162303:28
*** Yanyanhu has joined #senlin03:30
openstackgerritEthan Lynn proposed openstack/senlin: Add cluster lock steal  https://review.openstack.org/24618003:31
openstackgerritEthan Lynn proposed openstack/senlin: Add node_steal for node lock  https://review.openstack.org/24617903:31
*** lixinhui has quit IRC03:33
elynnYanyanhu, Qiming  Does 0.1.6 of senlinclient fix sdk problem?03:34
Qimingbetter use 0.1.8 I think03:35
YanyanhuI think 0.1.7 is workable03:35
Qimingthere were some other problems found later on03:35
Qiming"workable", :)03:35
Qimingno such word, Yanyanhu03:35
Yanyanhuok... so just 0.1.7 works03:35
elynnI thought 0.1.6 is the latest one, I made a mistake..03:36
openstackgerritMerged openstack/senlin: Add node_steal for node lock  https://review.openstack.org/24617903:51
*** elynn has quit IRC03:52
*** yuanying has joined #senlin04:07
Yanyanhuhi, Qiming, you may need to update the raises decription in this patch https://review.openstack.org/#/c/242681/4..6/openstack/connection.py04:07
Yanyanhusince three types of exception could be raises now04:08
openstackgerritMerged openstack/senlin: Fix documentation for conformance  https://review.openstack.org/25123004:10
Qimingem ... there are more than 3 types04:11
openstackgerritMerged openstack/senlin: Add cluster lock steal  https://review.openstack.org/24618004:12
Yanyanhuah, yes, my mistake, actually http exception will be raised04:12
openstackgerritjunxu proposed openstack/senlin: Store the actual UUID of obj_id in Senlin database for webhook create  https://review.openstack.org/25118704:56
*** elynn has joined #senlin05:00
*** elynn has quit IRC05:07
*** elynn has joined #senlin05:28
*** elynn has quit IRC06:03
*** elynn has joined #senlin06:04
xuhaiweiAnyone has the same problem? when requesting to Nova by creating/deleting a node, will get this error : public endpoint for compute service named senlin not found06:15
xuhaiweino context06:16
xuhaiweijust creating a node06:16
Qimingare you using the latest senlinclient?06:16
xuhaiweiyes, I think06:17
xuhaiweithe endpoint_filter fix has been pulled06:17
Qimingsomething is wrong, there is no "compute service named senlin"06:18
xuhaiweithe EndpointNotFound exception is raised from keystoneauth106:19
Qimingwe are requesting a "compute" service, named "senlin"06:19
*** pm90_ has joined #senlin06:21
xuhaiweiwhy the 'senlin' endpoint is not found06:21
Qimingthe error says "compute service named senlin"06:22
Qiming"compute service named senlin"06:22
Qimingread it again?06:22
xuhaiweishould be clustering service named senlin?06:22
Qimingyou are not creating or using webhooks?06:22
Qimingyes, senlin is not a compute service06:23
Qimingno compute service is named senlin06:23
Qimingsomehow, we are doing something weird06:23
Qimingfor example, in keystone_v3 service_get, why are we specifying both service_type and name?06:24
Qimingit could be the sdk side is broken as well06:25
xuhaiweiI have investiged that service_get method some time, it seems service_type and name are not correctly used in keystoneauth106:25
Qimingafter migrating to use keystoneauth1, the identity service needs a revision?06:25
xuhaiweithere are several get_endpoint method in keystoneauth1, don't know whether we can use one of them directly06:26
Qimingwell ... I don't think so06:27
Qimingit is sdk's problem to solve06:28
Qimingall we (senlin) should care about is the sdk interface06:28
xuhaiweisometime just dont know who is wrong, them or us06:29
Qimingright, that is a huge pain06:29
Qiming"when requesting to Nova by creating/deleting a node,"06:45
Qimingxuhaiwei, are you using senlin to do node creation?06:45
xuhaiweiyes, of course06:45
openstackgerritQiming Teng proposed openstack/senlin: Attempt to log more info on sdk exceptions  https://review.openstack.org/25167906:47
Qimingxuhaiwei, try this patch ^06:48
Qimingsee this can help give you a more detailed context06:48
QimingI'm moving the LOG.exception so that it is in an exception context06:49
Qimingthe previous location where we do LOG.exception was not working, because it is already outside an exception context06:50
xuhaiweiyes, much more details now06:51
Qimingwe have been logging a lot of info before, until it broke the fixture06:54
*** xuhaiwei has quit IRC06:56
*** xuhaiwei has joined #senlin06:56
xuhaiweiQiming, I noticed we supported nested cluster in senlin07:24
xuhaiweibut in fact, this design is not well used currently, is there any use case for it?07:25
Qimingwe supported?07:26
xuhaiweiwe are supporting07:26
Qimingnot sure we are supporting07:27
Qimingwe don't have a overall design for that07:27
xuhaiweiwe create a cluster, we can configure its parent07:27
Qimingthe 'parent' parameter is just a placeholder07:27
xuhaiweican it be used in some case?07:28
Qimingyes I think so07:28
xuhaiweior in the future07:28
Qimingbut I am not sure what it would be07:28
xuhaiweito me, when saying a cluster has a son, I may want to caculate all the nodes to its current_size07:30
xuhaiweithat's wrong, right07:30
Qimingquestions like that all need answers07:31
QimingI don't think we are gonna support that in Mitaka07:31
xuhaiweiin fact i dont want to support it in senlin, that will totally break the current design I think07:32
Qimingthere will be rules07:32
*** pm90_ has quit IRC07:34
xuhaiweianother question, about the senlinclient/v1/shell.py test, I am not testing the 'args', is it ok? Or I need to add them or something else07:35
Qimingnot testing args ... what do you mean?07:37
xuhaiweiI mean the 'options'07:39
xuhaiweiwhich is required or not07:39
Qimingpoint me to an exact line number?07:50
xuhaiweisomething like this07:52
Qimingit depends07:57
xuhaiweiactually I am not clear how to test them07:57
Qimingmaybe check this? https://github.com/openstack/python-senlinclient/blob/master/senlinclient/tests/unit/test_shell.py07:58
Qimingor more specifically, this function: https://github.com/openstack/python-senlinclient/blob/master/senlinclient/tests/unit/test_shell.py#L16907:59
xuhaiweiI will think about it08:05
Qimingokay, thx08:07
openstackgerritMerged openstack/senlin: Attempt to log more info on sdk exceptions  https://review.openstack.org/25167908:07
Qimingabout your patch on 202, I've blocked it08:07
QimingI mean this one: https://review.openstack.org/#/c/251669/08:08
xuhaiweiok, you mean we should check the 'action' one by one?08:09
Qiminglet's hold it, I'm proposing a different approach, let's see my proposal is a better option08:09
Qimingthe only confusion I have is about the DELETE request08:11
xuhaiweiwhat do you mean08:13
Qimingsee this: http://specs.openstack.org/openstack/api-wg/guidelines/http.html08:13
Qimingfind section "2xx success codes"08:14
Qimingit says this: Synchronous resource deletion08:14
QimingResponse status code must be 204 No Content08:14
QimingThere is no guidance on asynchronous resource deletion08:15
Qimingbut in real life, most resource deletion would be done asynchronously, say nova server deletion, heat stack deletion, senlin cluster deletion08:16
Qimingsome deletions may take quite some time to complete08:16
xuhaiweiit says 'all others should be 200'08:16
xuhaiweiincluding Asynchronous resource deletion ?08:17
Qimingthat is a gray area08:19
Qimingmy understanding is that 202 might be more appropriate for asynchronous deletion08:26
Qimingsee http://tools.ietf.org/html/rfc2616#section-10.2.308:26
xuhaiweiit's difficult to understand08:29
Qimingbut if you compare it to 10.2.1 (200), 20.2.2 (201), 202 is more appropriate08:30
xuhaiweito me, I think all asynchronous actions should return 202, no matter 'creation' or 'deletion'08:32
Qimingyep, agreed08:32
Qimingneither creation nor deletion has been completed when we return08:32
Qiming201 for creation is only suitable for profile, policy08:33
xuhaiweiso in senlin, we should filter out 'asynchronous actions' first?08:33
Qiming204 for deletion is only suitable for profile, policy, etc. not clusters or nodes08:33
xuhaiweiin senlin all the 'actions' are asynchronous, right?08:34
QimingI'm walking through all the apis08:34
Qimingyou cannot say that08:34
xuhaiweiin the api layer, it maybe not08:34
*** pm90_ has joined #senlin08:35
Qimingthe 'actions' in senlin's terminology are all asynchronously executed08:35
Qimingregarding the 'action' defined in the 'routes' package, it is different08:35
*** pm90_ has quit IRC08:39
xuhaiweiwhat's that? I can't open the url08:58
Yanyanhulooks like a huge diagram of http request handling workflow...09:04
xuhaiweimy company's security system blocked it09:05
openstackgerritxu-haiwei proposed openstack/python-senlinclient: Fix help message error for node_create  https://review.openstack.org/25173009:25
Qimingxuhaiwei, sent you via email09:25
*** yuanying has quit IRC09:30
xuhaiweioh my eyes !!09:30
xuhaiweihow is your nose today? :)09:31
Yanyanhuvery uncomfortable09:35
Yanyanhuand also my throat and lung...09:35
xuhaiweiwearing musk?09:35
Yanyanhuyou even can't see a building which is 100 meters away09:36
Yanyanhuyes, got one from the company09:36
Yanyanhuwill wear it when getting back to home09:36
xuhaiweimaybe need to ware two at one time09:36
Yanyanhuyea :P09:36
*** Yanyanhu has quit IRC09:52
*** zhenguo has quit IRC10:04
*** pm90_ has joined #senlin10:15
*** elynn has quit IRC10:22
*** pm90_ has quit IRC10:23
openstackgerritQiming Teng proposed openstack/senlin: Fix status code returned from API  https://review.openstack.org/25177110:41
Qimingxuhaiwei, still online?10:42
Qimingthe above patch is the one I mentioned10:42
Qimingplease help review10:42
Qimingwe have got to do this one by one10:42
Qimingnot very urgent10:57
openstackgerritxu-haiwei proposed openstack/python-senlinclient: Add test case for v1/shell.py part2  https://review.openstack.org/25179811:15
*** lixinhui has joined #senlin11:41
lixinhuihow for me to touch this guy11:46
*** elynn has joined #senlin12:20
*** pm90_ has joined #senlin12:27
*** yanyanhu has joined #senlin12:30
*** jruano has joined #senlin13:04
*** pm90_ has quit IRC13:15
*** yanyanhu has quit IRC14:00
*** haiwei has joined #senlin14:00
*** elynn has quit IRC14:00
haiweianything else?14:00
Qimingnope from me sir14:01
Qimingit is too late for you14:01
Qimingtime to date someone in dream14:01
haiweidon't be too tired14:01
haiweiyou too14:01
haiweisee you14:02
Qimingsee you14:02
*** haiwei has quit IRC14:06
*** lixinhui has quit IRC14:11
*** zhenguo has joined #senlin14:18
*** Qiming has quit IRC14:35
*** zhenguo has quit IRC15:05
*** pm90_ has joined #senlin15:20
*** pm90__ has joined #senlin15:23
*** pm90_ has quit IRC15:25
*** jruano has quit IRC16:12
*** pm90__ has quit IRC17:43
*** pm90_ has joined #senlin19:15
*** pm90_ has quit IRC23:01
*** jruano has joined #senlin23:46

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