Wednesday, 2016-12-07

*** yuanying_ has joined #senlin00:03
*** yuanying has quit IRC00:06
*** yuanying_ has quit IRC00:07
*** elynn has joined #senlin00:09
*** elynn has quit IRC00:21
*** catintheroof has quit IRC00:23
*** catintheroof has joined #senlin00:24
*** catintheroof has quit IRC00:28
*** elynn has joined #senlin00:31
*** devlaps has quit IRC00:37
*** devlaps has joined #senlin00:38
*** yuanying has joined #senlin01:04
*** dinghh has quit IRC01:07
*** dinghh has joined #senlin01:12
*** zhurong has joined #senlin01:14
*** dinghh has quit IRC01:16
*** guoshan has joined #senlin01:24
*** dinghh has joined #senlin01:24
*** zhurong has quit IRC01:25
*** zhurong has joined #senlin01:26
*** dinghh has quit IRC01:30
*** zhurong has quit IRC01:34
*** zhurong has joined #senlin01:35
*** chohoor has joined #senlin01:37
*** fabian4 has quit IRC01:40
*** dinghh has joined #senlin01:43
*** fabian4 has joined #senlin01:43
*** yanyanhu has joined #senlin01:48
openstackgerritMerged openstack/senlin-dashboard: Following patch for angular cluster table  https://review.openstack.org/40732301:51
*** shu-mutou-AWAY is now known as shu-mutou02:01
openstackgerritAaron Ding proposed openstack/senlin: Correct the installation doc, fix path of openrc  https://review.openstack.org/40748502:03
*** lxinhui has joined #senlin02:04
openstackgerritRUIJIE YUAN proposed openstack/senlin: fields object support webhook trigger  https://review.openstack.org/40780602:19
*** guoshan has quit IRC02:31
*** guoshan has joined #senlin02:32
*** elynn has quit IRC02:46
*** elynn has joined #senlin02:48
*** yuanying has quit IRC02:58
*** guoshan has quit IRC03:06
*** guoshan has joined #senlin03:07
openstackgerritRUIJIE YUAN proposed openstack/senlin: engine service support webhook-trigger2  https://review.openstack.org/40783203:33
ruijiehttp://logs.openstack.org/15/407515/1/check/gate-senlin-dsvm-tempest-integration-nv/f9037fb/console.html#_2016-12-06_14_21_47_88532403:34
ruijiethe gate was broken?03:34
*** Drago1 has quit IRC03:37
openstackgerritMerged openstack/senlin: Get both event drivers into action  https://review.openstack.org/40659503:42
*** yuanying has joined #senlin03:43
Qimingruijie, could be this https://review.openstack.org/#/c/407274/103:46
*** yuanying has quit IRC03:48
*** yuanying has joined #senlin03:49
*** yuanying has quit IRC03:53
*** chohoor has quit IRC03:53
*** chohoor has joined #senlin03:54
*** xuhaiwei has joined #senlin03:54
*** devlaps has quit IRC03:58
*** guoshan has quit IRC04:03
*** zhurong has quit IRC04:27
*** yuanying has joined #senlin04:40
*** elynn has quit IRC05:03
*** guoshan has joined #senlin05:04
*** guoshan has quit IRC05:08
*** yuanying has quit IRC05:13
*** yuanying has joined #senlin05:13
*** zhenguo_ has joined #senlin05:14
*** zhenguo has quit IRC05:15
*** zhenguo_ is now known as zhenguo05:15
*** zhurong has joined #senlin05:27
*** zhurong has quit IRC05:30
*** zhurong has joined #senlin05:31
yanyanhuhi, lxinhui, around?05:43
lxinhuihi, yanyanhu05:47
lxinhuiI am here05:47
lxinhuimy laptop crashed in the weekend for new VMware VPN, jus recover everything05:49
yanyanhulxinhui, sigh...05:57
yanyanhusorry just talked with a colleague05:58
yanyanhuabout this patch, https://review.openstack.org/#/c/402296/05:58
yanyanhuI think it looks good. Just I have the same question as Michael about the location that event is generated, I guess we should emit event in "update_health" method?06:00
lxinhuiAccording to the spec06:02
lxinhuiof lbaasv206:02
lxinhuiall the details of different member of lbaas pool will be reported by stats06:03
yanyanhulxinhui, you mean health status is just a subset of stats?06:04
*** guoshan has joined #senlin06:04
lxinhuihealth is different API06:05
lxinhuiwhat we want is the details of different member ndoe06:05
lxinhuisuch as alive, dead, or reconfiged06:05
lxinhuiby reading the code, you will find nowdays the method of udate_db will do the rpc call06:06
lxinhuiof neutron_lbaas to update the db with the latest details collected from lbaas listener06:06
lxinhuiwhat we want is to get such information directly instead of read DB06:07
yanyanhuyes, that makes sense06:07
yanyanhujust I noticed both UpdateStatsDb and UpdateHealthDb define a method "emit"06:08
yanyanhuso I was wondering if we only care about health status, maybe we should event_notifier.emit in UpdateHealthDB.emit rather than UpdateStatsDB.emit06:08
*** guoshan has quit IRC06:09
yanyanhuactually I'm not very clear about the difference between this health status and stats :)06:09
lxinhuiYes06:09
yanyanhubased on the comment embedded in the code06:10
yanyanhuI guess health status is included in stats06:10
lxinhuithat depends on who will call udpdatehalthDB06:10
yanyanhuand stats contains more information like "ereq", "conns", etc.06:10
yanyanhubut health is only about status, e.g. UP, ONLINE, etc. of pool, member, and listener06:11
yanyanhulxinhui, you mean?06:12
lxinhuiSorry I have a meeting06:12
lxinhuiwill ping you after that06:12
yanyanhusure06:12
yanyanhuthanks a lot :)06:12
*** guoshan has joined #senlin06:24
*** lxinhui_ has joined #senlin06:30
*** lxinhui has quit IRC06:34
*** zhurong has quit IRC07:01
*** zhurong has joined #senlin07:02
*** yuanying has quit IRC07:05
*** yuanying has joined #senlin07:17
*** edisonxiang has joined #senlin07:39
edisonxiangQiming: hello07:39
Qiminghi07:39
edisonxiangQiming: sorry to trouble you. Could you tell me how to publish the Senlin's API-ref?07:43
Qimingit is already published07:44
edisonxiangyeah.07:44
Qiminghttp://developer.openstack.org/api-ref/clustering/07:44
edisonxiangcould you tell me how to do that?07:44
edisonxiangthanks07:44
Qimingit is a gate job07:46
edisonxiangOK. thanks07:48
Qiminglemme find the patch07:49
edisonxiang:)07:49
edisonxiangyou are so nice07:49
Qiminga long history ...07:53
Qiminghttps://review.openstack.org/#/c/314832/07:53
Qimingfinally found it07:53
edisonxianggreat07:57
edisonxiangthanks Qiming07:57
*** zhurong has quit IRC08:00
*** zhurong has joined #senlin08:01
Qimingnp08:02
*** yuanying has quit IRC08:08
edisonxiang:)08:17
lxinhui_yanyanhu08:19
lxinhui_still thereļ¼Ÿ08:19
Qimingsaw him alive, :D08:25
Qimingvivid08:25
ruijie:)08:32
openstackgerritKenji Ishii proposed openstack/senlin-dashboard: Add create cluster action  https://review.openstack.org/40347208:35
openstackgerritQiming Teng proposed openstack/senlin: Remove 'InvalidParameter' exception type  https://review.openstack.org/40793208:36
yanyanhuhi, lxinhui_ sorry just noticed your message08:48
yanyanhuhi, xuhaiwei, around?08:52
xuhaiweiyes08:53
yanyanhuhi, about the proposal to Tacker08:54
openstackgerritMerged openstack/senlin-dashboard: Add create cluster action  https://review.openstack.org/40347208:54
xuhaiweiyes?08:54
yanyanhumaybe we can have a quick sync before talking to Tacker team?08:54
yanyanhuI assume we will join their next weekly meeting in Dec.14 :)08:54
xuhaiweiI have summarized it in this spec https://review.openstack.org/#/c/352943/08:55
yanyanhuyes, I believe all the team are interested on it :008:55
xuhaiweibriefly speaking, We will make Senlin as a driver for vnf scaling08:56
yanyanhubut some ones may still not quite understand the detail08:56
yanyanhuyes, that's the goal08:56
xuhaiweiyanyanhu, shall we introduce senlin again to the team?08:56
openstackgerritQiming Teng proposed openstack/senlin: Remove 'ProfileTypeNotMatch' exception type  https://review.openstack.org/40793808:57
yanyanhuxuhaiwei, umm, maybe not, unless some ones from Tacker team want to listen it08:57
yanyanhuxuhaiwei, maybe we can make a quick call on Thursday or Friday? if it's convenient for you08:57
Qimingoh, haiwei, we have done that back in Barcelona08:58
xuhaiweiyes, I just want to share the idea of how to integrating senlin to tacker08:58
Qimingwe tried very hard to educate Tacker team on the scope, capability and status of Senlin08:58
Qimingagree that a discussion to get everyone onto the same page is important08:59
*** yuanying has joined #senlin08:59
xuhaiweiyes, Qiming, so I think we dont need to introduce what is senlin again08:59
QimingI don't think there is a need08:59
yanyanhuyes, so maybe we can sync up what we want to express in that meeting08:59
xuhaiweijust talk about how to integrate08:59
yanyanhuto better convince them08:59
yanyanhuand also make Senlin team better understand Tacker :)09:00
Qimingright, it is a IRC meeting, no time to walk everyone through senlin09:00
Qimingyes, good point09:00
Qimingwe need to listen and understand what their concerns09:00
yanyanhuyes09:00
xuhaiweiyanyanhu, if you read the spec, you can understand what is going to be done09:00
Qiminghaiwei, personally, I'm not that worried what could be done, I'm more interested in what should be done09:01
xuhaiweigenerally speaking, the goal should be the scalability and HA of vnf09:01
QimingI mean, what are their requirements, explicitly or implicitly expressed ...09:02
yanyanhuxuhaiwei, yes, I have read it. Just still not quite clear about the translation part :) I guess I need to learn some basic conception about VDU09:02
yanyanhuand also the translation of TOSCA09:02
xuhaiweithe scalability and HA can be controlled by senlin scaling-policy and ha-policy09:03
xuhaiweiI think that's why senlin meets Tacker's need09:03
Qimingthere are basically two tech solutions: 1. have translator translates everything into Heat template, then tacker invokes Senlin indirectly as a resource type, that is already doable, according to Sahdev, PTL of tosca parser09:03
Qiming2. have tosca parser translate high-level specifications (HA, scaling, etc) into Senlin policy specs directly, then Tacker interacts with Senlin directly09:04
QimingI'm more opt in for the 2nd approach because Tacker will really benefit from the rich APIs Senlin exposes09:05
xuhaiweiyanyanhu, the translation part needs some jobs then, currently heat-translator has supported translation of senlin scaling policy, but tacker team doesnt allow new resources in their template, I have to translate the template in Tacker again instead of using heat-translator directly09:05
Qimingbut I understand Tacker team's concern about the feasiblity of solution 2, they don't know that translation is doable or not, or anyone will help get that done09:06
Qiminganother implicit concern could be that Tacker wants to manage their dependencies to other services to a minimum09:06
yanyanhuxuhaiwei, I see. So they prefer to manage all resources (including senlin cluster) using Heat in current stage09:07
yanyanhuincluding senlin cluster/policy09:07
xuhaiweiQiming, exactly as you said, just one thing need to mention, that is Tacker already has their scaling policy, so senlin's policy translation by heat-translator can not be used directly09:07
Qiming"doesn't allow new resources in their template", you mean the high-level TOSCA specification?09:07
Qimingat which level Tacker's scaling policy sits? is it above Heat template, or below that?09:08
xuhaiweiQiming, yes, Tacker defined their own scaling policy in TOSCA template already09:08
xuhaiweiabove it09:08
Qimingyanyanhu, I think the team there is still struggling09:08
Qimingxuhaiwei, if it is above it, it is irrelevant09:08
Qimingit will be translated to heat policy resource type or senlin policy spec anyway09:09
xuhaiweiQiming, they are not struggling I think09:09
Qimingokay, then they will be regretting09:09
xuhaiweiQiming, that is not a big problem, senlin scaling policy can be under tacker scaling policy type, it just a policy name imo09:10
Qimingthe key point is ... Heat is not designed to be a tool for operators09:10
yanyanhuso, that's why we need to translate Tocsa template to Heat template first, then use it to define Senlin profile.09:10
yanyanhusorry, s/we/they :)09:10
Qimingexactly, xuhaiwei, so we don't have to express the 'senlin' name in their high-level policy spec09:10
xuhaiweiyes, we dont need to09:10
Qimingjust when tacker turns those definitions into something real, it can be translated to heat ones or senlin ones09:11
Qimingwe won't be able to convince them away from heat09:11
xuhaiweiQiming, you are right09:11
xuhaiweiin Tacker all the things are done by heat, including senlin09:11
Qimingyanyanhu, maybe I'm too eager to see them succeed09:12
yanyanhu:)09:12
xuhaiweisenlin is just managing their vnfs  after vnf are created09:12
Qimingokay, they will realize, soon, they they will change their deployment09:12
yanyanhuwe both, for that means we succeed as well if they use senlin to manage their cluster :)09:12
Qimingand the only way to do that, for them, is 'heat stack-update'09:12
Qimingsenlin is just a tool09:13
Qimingit is not designed for a specific business use case09:13
Qimingso ... we'd better prepared for that, :)09:13
*** dinghh1 has joined #senlin09:14
*** dinghh has quit IRC09:14
Qimingthat is why I am so eager to see tacker succeed, magnum succeed, zun succeed, megh...war succeed09:14
yanyanhuhaha09:15
xuhaiweiQiming, yanyanhu, though senlin scaling policy can not be added in Tosca template directly, other senlin policies can be added directly, that means we should support these policy translation in heat-translator09:15
Qimingthey succeeded because they are using senlin for resource pool management, ... that's my dream09:15
QimingTOSCA templates already hardcoded heat policy?09:15
yanyanhuxuhaiwei, that is a work in heat-translator?09:15
QimingI believe all other policies, including scaling policy, specification was derived from senlin09:16
Qimingwe helped shape that standard09:16
xuhaiweiQiming, yes, Tacker already difined their scaling policy09:16
Qimingin TOSCA?09:16
xuhaiweiyes09:16
Qimingthen why we want to change that for senlin's purpose?09:17
xuhaiweiTacker's scaling policy is derived from TOSCA scaling policy09:17
Qimingit is not generic enough?09:17
Qimingokay, that is fine09:17
Qimingthat means it can be translated to either heat or senlin scaling policy09:17
Qimingjust that in senlin's design, we don't include the "alarm" part09:18
xuhaiweisenlin scaling policy is also derived from TOSCA scaling policy, so senlin's policy is conflicting with Tacker's one09:18
Qimingwe felt that inclusion a stupid constraint09:18
Qimingxuhaiwei, that is where I was confused09:18
Qiming"Tacker's scaling policy is derived from TOSCA scaling policy"09:19
Qiming"senlin scaling policy is also derived from TOSCA scaling policy, so senlin's policy is conflicting with Tacker's one"09:19
Qimingwhy?09:19
xuhaiweimaybe TOSCA allows many types of scaling policies09:19
Qimingthey should do that IMO09:19
Qimingdo you mean there is a workflow like this?09:19
QimingTOSCAL scaling policy -> Tacker scaling policy -> Heat Scaling policy?09:20
yanyanhuI guess not?09:20
xuhaiweiTacker scaling policy = Heat scaling policy imo09:20
Qimingthat was my guess09:20
xuhaiweijust the type is changed09:21
yanyanhuIMO, it should be TOSCA scaling policy -> Heat Scaling policy  or TOSCA scaling policy -> Senlin scaling policy09:21
Qimingso we are not talking about tacker's scaling policy09:21
xuhaiweiyes09:21
xuhaiweiyou both are right09:21
Qimingit is about two translations: TOSCAL->Heat, TOSCAL-Senlin09:21
Qimingyes09:21
Qimingthen there is no conflict09:21
Qimingthe 2nd translation is not there yet, but we can get help from the translator team easily09:22
Qimingit is trivial09:22
xuhaiweibut in Tacker every resources are Tacker types09:22
Qimingthat is fine09:23
Qimingunder their roof, follow their design09:23
Qimingproblem is?09:23
xuhaiweino problem09:23
xuhaiweijust follow tacker's scaling policy type, make senlin as a driver, that will be ok09:24
Qimingexactly ...09:24
xuhaiweisenlin scaling policy and heat scaling policy are two drivers of tacker scaling policy09:24
yanyanhuI see09:24
yanyanhuit's more clear for me now :)09:24
Qimingfahongbao09:25
xuhaiweihope tacker team will feel happy about that :)09:25
yanyanhu-_-!09:25
yanyanhuyes, hope so09:25
yanyanhuso lets push it to move on step by step09:25
xuhaiweithe source code is not difficult i think, I have started this job by refactoring the existing code in Tacker09:26
yanyanhuI believe once they get better understanding of Senlin, they will try to use it to support more functionality09:27
yanyanhugreat09:27
yanyanhuwe will also join that meeting in Dec.14, and help you answer their question :)09:27
yanyanhuand also listen :P09:27
xuhaiweithanks yanyanhu09:27
yanyanhumy pleasure09:28
openstackgerritRUIJIE YUAN proposed openstack/senlin: fields object support webhook trigger  https://review.openstack.org/40780609:41
openstackgerritRUIJIE YUAN proposed openstack/senlin: engine service support webhook-trigger2  https://review.openstack.org/40783209:41
openstackgerritRUIJIE YUAN proposed openstack/senlin: api support webhook-trigger2  https://review.openstack.org/40797109:41
openstackgerritRUIJIE YUAN proposed openstack/senlin: kill dead code in engine service  https://review.openstack.org/40797309:44
openstackgerritRUIJIE YUAN proposed openstack/senlin: kill dead code in engine service and rpc  https://review.openstack.org/40797309:46
*** zhurong has quit IRC10:03
*** guoshan has quit IRC10:42
*** yanyanhu has quit IRC10:46
*** openstackgerrit has quit IRC11:32
*** openstackgerrit has joined #senlin11:42
*** ChanServ sets mode: +v openstackgerrit11:42
openstackgerritShu Muto proposed openstack/senlin-dashboard: Add node create/update actions  https://review.openstack.org/39697511:42
*** guoshan has joined #senlin11:43
*** shu-mutou is now known as shu-mutou-AWAY11:47
*** guoshan has quit IRC11:47
*** catintheroof has joined #senlin11:55
*** chohoor has quit IRC11:58
*** catintheroof has quit IRC12:32
*** catintheroof has joined #senlin12:32
*** catintheroof has quit IRC12:37
*** guoshan has joined #senlin12:44
*** guoshan has quit IRC12:48
*** catintheroof has joined #senlin12:54
*** guoshan has joined #senlin13:04
*** catinthe_ has joined #senlin13:14
*** guoshan has quit IRC13:15
*** catintheroof has quit IRC13:16
*** lxinhui_ has quit IRC13:28
openstackgerritOpenStack Proposal Bot proposed openstack/senlin: Updated from global requirements  https://review.openstack.org/40811413:40
*** catintheroof has joined #senlin14:02
*** catinthe_ has quit IRC14:04
openstackgerritMerged openstack/senlin: Fix profile create - call subclass methods  https://review.openstack.org/40740114:28
*** guoshan has joined #senlin15:17
*** guoshan has quit IRC15:21
*** Drago1 has joined #senlin15:23
*** catintheroof has quit IRC15:26
*** catintheroof has joined #senlin15:26
*** catintheroof has quit IRC15:31
*** catintheroof has joined #senlin15:57
*** guoshan has joined #senlin16:18
*** guoshan has quit IRC16:22
*** devlaps has joined #senlin16:43
*** guoshan has joined #senlin17:18
*** guoshan has quit IRC17:23
*** dinghh has joined #senlin17:41
*** dinghh1 has quit IRC17:42
*** guoshan has joined #senlin18:19
*** guoshan has quit IRC18:24
*** guoshan has joined #senlin20:21
*** guoshan has quit IRC20:25
*** guoshan has joined #senlin21:21
*** guoshan has quit IRC21:26
openstackgerritOpenStack Proposal Bot proposed openstack/senlin: Updated from global requirements  https://review.openstack.org/40811421:31
*** catintheroof has quit IRC21:37
*** catintheroof has joined #senlin21:38
*** catintheroof has quit IRC21:42
*** guoshan has joined #senlin22:22
*** guoshan has quit IRC22:27
*** guoshan has joined #senlin23:23
*** guoshan has quit IRC23:27

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