Wednesday, 2015-09-09

« Tuesday, 2015-09-08 Index Thursday, 2015-09-10 »
*** Qiming__ has quit IRC00:08
*** jdandrea has quit IRC00:58
*** zhenguo has joined #senlin01:02
*** Qiming__ has joined #senlin01:09
*** admin__ has quit IRC01:10
Qiming__morning, :)01:11
*** Shijia has joined #senlin01:13
*** yonglihe has joined #senlin01:15
*** Qiming is now known as Qiming-mobile01:16
*** Qiming__ is now known as Qiming01:16
*** xuhaiwei has joined #senlin01:22
openstackgerritCindia-blue proposed stackforge/senlin: Add a placement-policy to enable vSphere DRS functions  https://review.openstack.org/21921201:26
*** lixinhui has joined #senlin01:27
*** Yanyanhu has joined #senlin02:06
*** elynn has joined #senlin02:19
*** lixinhui has quit IRC02:30
*** lixinhui has joined #senlin02:36
*** Shijia has quit IRC03:24
*** Shijia has joined #senlin03:25
*** Shijia has quit IRC03:25
*** Shijia has joined #senlin03:25
*** Shijia has quit IRC03:48
*** Shijia has joined #senlin03:49
*** Shijia has quit IRC04:07
*** Shijia has joined #senlin04:08
*** Shijia has quit IRC04:12
*** Shijia has joined #senlin04:18
*** Shijia has quit IRC04:22
*** Shijia has joined #senlin04:24
*** Shijia has quit IRC04:27
*** Shijia has joined #senlin04:27
*** Shijia has quit IRC04:34
*** Qiming-mobile has quit IRC04:56
*** elynn has quit IRC06:01
*** elynn has joined #senlin06:08
openstackgerritYanyan Hu proposed stackforge/senlin: Fix a bug in cluster scale in action  https://review.openstack.org/21874006:21
*** branw has joined #senlin06:35
branwwhen i curl -i -X http://webhookurl, i got 403 Forbidden, http://paste.openstack.org/show/451504/06:38
xuhaiweibranw, maybe your policy.json is old06:39
xuhaiweicheck /etc/senlin/policy.json the fisrt line, see if you are using project_id instead of 'project'06:40
branwthere is no project or project_id in the first line06:42
xuhaiwei"admin_or_owner":  "is_admin:True or project:%(project)s",06:43
xuhaiweisorry, the second line06:43
branwi got the  "admin_or_owner":  "is_admin:True or project:%(project)s",06:43
xuhaiweimaybe for the webhook trigger, can only use admin user? not sure about it06:46
xuhaiweiadmin user is ok?06:47
xuhaiweiI saw your command, you are using -c06:48
branwthe webhooks.rst in getting_started  use senlin webhook-create -c test-cluster -a CLUSTER_SCALE_OUT -C 'user=$USER_ID;password=$PASSWORD' test-webhook06:50
xuhaiweiit's strange06:56
Qimingbranw, you will need to add '-d' to the senlin command line to diagnose how request is sent to senlin-api07:00
Qimingon the other hand, you will need to check the log from senlin-api and senlin-engine to understand where the 403 exception was thrown07:01
openstackgerritYanyan Hu proposed stackforge/senlin: Fix a bug in cluster scale in action  https://review.openstack.org/21874007:31
*** elynn_ has joined #senlin08:08
*** elynn has quit IRC08:08
openstackgerritYanyan Hu proposed stackforge/senlin: Fix a bug in cluster scale in action  https://review.openstack.org/21874008:24
*** elynn_ has quit IRC08:40
*** elynn_ has joined #senlin08:41
openstackgerritMerged stackforge/senlin: Fix a bug in cluster scale in action  https://review.openstack.org/21874008:46
Qiminghi, guys09:05
Qiminghttp://lists.openstack.org/pipermail/openstack-dev/2015-September/073897.html09:05
Qiminghave you noticed this?09:06
xuhaiweiwe may use aodh instead of ceilometer?09:07
Yanyanhuhave read it. So the change has been formally made?09:07
Qimingalso this one: http://lists.openstack.org/pipermail/openstack-dev/2015-September/073793.html09:08
YanyanhuGnocchi is also splitted09:09
Yanyanhuwhat is this project for?09:09
YanyanhuTDBaaS (Time Series Database as a Service)09:12
Yanyanhuis it in the scope of Trove?09:13
Qimingno09:17
QimingIf I'm understanding this correctly, Gnocchi is only used for time series processing, not generic interface of DB09:18
Yanyanhuok09:19
QimingI'm replying that email for some clarifications09:19
*** elynn_ has quit IRC09:19
Yanyanhuso it's a layer upon db09:19
Qimingreading that email, any questions you want to ask?09:23
Yanyanhuthe first one?09:24
Qimingyep09:24
Qimingbetter asking questions in single email09:24
*** elynn_ has joined #senlin09:25
Yanyanhuwhat I'm wandering is the change about API interface09:26
Yanyanhusince they promised to keep client compatible with old version09:27
Yanyanhubut seems user need to handle the interaction between ceilometer/Aodh/Gnocchi by themselves now09:28
Yanyanhuif they are using API interface directly09:29
Qimingclient and api support has been added to common Ceilometer interfaces09:29
Qimingsuch that if Aodh is enabled, the client can still be used and redirect09:29
Qimingto Aodh.09:29
Yanyanhuyes, but currently, user don't need to consider where is the source of metric when creating an alarm09:30
Yanyanhubut after this changes, seems they have to specify where the metric comes from explicitly?09:31
Yanyanhue.g. from ceilometer service in Regionxxx09:31
YanyanhuI guess Gnocchi won't influence the user interface09:31
QimingGnocchi is more of a library than a service09:32
Yanyanhuyes, seems so09:33
Qimingan alarm still has rules09:33
Qimingand those rules will determine which metrics will be used09:33
Yanyanhuyes, and I guess it will have a param(or config option?) to define where are those metric data it will evaluate09:34
Qimingconfig what?09:35
Yanyanhuso the interaction beween ceilometer and Aodh is API interface?09:35
Yanyanhuthrough rest API09:35
Yanyanhuconfig where is the ceilometer service09:35
Qimingthe mail didn't mention whether Aodh needs a separate endpoint09:35
Qimingceilometer can always be configured09:36
Qimingso the real question is whether Aodh is a standalone service, right?09:37
Yanyanhuyes, so theoretically Aodh in RegionOne can evaluate an alarm based on the performance data from both RegionOne and RegionTwo I guess :)09:37
Yanyanhuright09:37
Yanyanhuif it is, it can do what I said09:37
Qimingif it is not09:37
Qimingit is deployed with a ceilometer server09:38
Qimingin the same region09:38
Yanyanhuif not, I guess there is no big change09:38
Yanyanhuand Telemetry serivce(s) will still work as the same way I think09:39
Qimingyep09:39
Yanyanhuas it is now09:39
Qimingif the API doesn't change09:40
Qimingthis reorg is only a ceilometer-only change09:40
Yanyanhuyes09:41
Qimingguess what09:42
QimingI just posted the email and I now realized that I was asking some stupid questions09:42
Yanyanhu...09:42
Qimingcheck this: http://git.openstack.org/cgit/openstack/aodh/tree/devstack/settings09:43
Qimingaodh has its own APIs09:43
Yanyanhuan isolated service09:43
Qimingyep09:43
Qimingit has four processes at least09:44
Yanyanhuem09:44
Yanyanhujust those 'ceilometer-alarm-xxx' agents plus the api :)09:45
Qimingtrue is that aodh is reusing ceilometer client09:45
Qiminghttp://git.openstack.org/cgit/openstack/aodh/tree/devstack/plugin.sh#n24409:45
Qimingit is calling itself 'alarming' service09:46
Qiminghttp://git.openstack.org/cgit/openstack/aodh/tree/devstack/plugin.sh#n12209:46
Yanyanhuyea. I guess if they don't, people will be mad with them ;)09:46
Qimingnow I'm mad with them09:48
Yanyanhu:)09:49
xuhaiweiso shall we use it?09:49
Yanyanhuwe are using API not client09:49
Qimingyep09:49
Yanyanhuxuhaiwei, I guess we need to relearn the new API interface of aodh...09:49
Qimingwe need to add aodh resources to sdk09:49
Qimingand use them there09:50
xuhaiweiceilometer will remove alarm api?09:50
Yanyanhuyes09:50
Qimingthat part I'm not sure09:51
Yanyanhubut this is a problem since they will not accept new resource until their current work is done09:51
Qimingit won't happen anytime soon09:51
*** branw has quit IRC09:51
Yanyanhuyes09:51
Qimingso ...09:54
Qimingafter this clarification09:54
QimingI think any code based on ceilometer alarms would be a waste of time09:55
Yanyanhuyes09:55
Yanyanhuguess so09:55
Qimingespecially for us, due to the fact that we don't use xxxclient directly09:55
Qimingwe invoke their APIs thru openstacksdk09:55
xuhaiweiwho started aodh? does the ceilometer guys agree?09:57
Yanyanhuit should be a sub project splitted from ceilometer09:58
Yanyanhujust like lbaas and neutron09:58
openstackgerritYanyan Hu proposed stackforge/senlin: Rework context initialization in action module  https://review.openstack.org/22167610:03
*** elynn_ has quit IRC10:06
*** zhenguo has quit IRC10:11
openstackgerritLiuWei proposed stackforge/senlin: (WIP) cross-az policy  https://review.openstack.org/22168410:16
*** admin__ has joined #senlin10:16
*** admin__ has left #senlin10:19
*** R_lixh has joined #senlin10:19
openstackgerritYanyan Hu proposed stackforge/senlin: Add functional test for cluster scale in/out  https://review.openstack.org/21814210:32
*** Qiming has quit IRC10:34
*** Yanyanhu has quit IRC10:37
*** Qiming has joined #senlin11:29
*** lixinhui has quit IRC11:29
openstackgerritMerged stackforge/senlin: Add functional test for cluster scale in/out  https://review.openstack.org/21814211:42
openstackgerritMerged stackforge/senlin: Cluster action test case (4)  https://review.openstack.org/22117711:53
openstackgerritMerged stackforge/senlin: Revised SDK driver with unit test case  https://review.openstack.org/22046811:53
openstackgerritMerged stackforge/senlin: Cluster action test case (5)  https://review.openstack.org/22134312:47
*** zhenguo has joined #senlin12:59
*** jdandrea has joined #senlin13:27
*** jruano has joined #senlin15:14
*** Qiming has quit IRC15:50
*** jdandrea has quit IRC21:26
*** jruano has quit IRC23:05
*** xuhaiwei has quit IRC23:36
*** Qiming has joined #senlin23:47
« Tuesday, 2015-09-08 Index Thursday, 2015-09-10 »

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