Thursday, 2015-07-23

*** xuhaiwei has quit IRC00:03
openstackgerritMerged stackforge/senlin: Updated from global requirements
*** jruano has quit IRC00:36
*** lixinhui_ has joined #senlin00:37
*** xuhaiwei has joined #senlin00:42
*** hightall has joined #senlin00:48
*** Qiming_ has joined #senlin01:11
*** hightall has quit IRC01:31
*** jruano has joined #senlin01:32
*** mathspanda has joined #senlin01:36
xuhaiweiQiming_, in this patch, gate is always NOT_REGISTERED, do you know the reason01:40
Qiming_gerrit was down last night01:40
Qiming_it is now okay01:40
xuhaiweiI rechecked it, but nothing changed01:40
Qiming_-openstackstatus/#openstack-dev- NOTICE: CI is slowly recovering, please be patient while the backlog is worked through.01:41
Qiming_zuul has experienced some problems afict01:41
*** hightall has joined #senlin01:44
*** Yanyanhu has joined #senlin01:49
Qiming_lixinhui_, there?01:53
lixinhui_okay, thanks01:55
lixinhui_I will try it right now01:55
Yanyanhuhi, Qiming_ this patch When I tried to check pep8 error log, I got an error 'File Not Found'02:01
xuhaiweiI think this is also the gate's problem02:02
Qiming_just do a recheck?02:02
Yanyanhuok, let me have a try02:02
Yanyanhuxuhaiwei, me too02:02
Yanyanhuby the way, will start work on test case for lb_policy, please help to review :)02:03
*** Qiming_ is now known as Qiming02:16
lixinhui_Qiming_, the patch works02:22
lixinhui__get_create_body this method can be called when server new?02:34
Qimingit will be called automatically02:37
lixinhui_okay, thanks, Qiming02:38
Qimingthe weird thing is: scheduler_hints is only meant to be an out-of-band body entry for server creation02:39
Qimingyou won't be able to get it back when retrieving server details02:40
Qimingso ... I didn't add that into the Server class definition as a property02:40
lixinhui_en, I saw your revision to resource body when create. maybe because hints is designed for use of nova scheduler, but server-group, as a result of allocation/scheduling, will be part of server deails02:44
lixinhui_by the way, when openstacksdk to finish their release, the extension to server-group is also very useful to commit in02:45
Qimingyes, but we have to respect their decision to do a code freeze before release02:46
xuhaiweigate seems to be recovered02:47
*** Zhenqi has joined #senlin02:48
Qimingit was recovering late last night BJ time02:48
*** hightall has quit IRC03:19
Yanyanhuhi, Qiming, the cred attr in Credential DB model is a list of trusts?03:22
Yanyanhuor a single trust03:22
Qiminga single03:23
* Qiming is heading for lunch03:26
*** Qiming is now known as Qiming_afk03:26
*** mathspanda has quit IRC03:38
openstackgerritQiming Teng proposed stackforge/senlin: Remove some unused dependencies
*** Zhenqi has quit IRC04:09
*** Zhenqi has joined #senlin04:10
*** mathspanda has joined #senlin04:43
openstackgerritOpenStack Proposal Bot proposed stackforge/senlin: Updated from global requirements
*** Qiming_afk is now known as Qiming05:09
openstackgerritYanyan Hu proposed stackforge/senlin: Add test case for LoadBalancer driver(1)
openstackgerritMerged stackforge/senlin: Updated from global requirements
*** Tennyson has joined #senlin05:30
openstackgerritYanyan Hu proposed stackforge/senlin: Add test case for LoadBalancer driver(1)
openstackgerritYanyan Hu proposed stackforge/senlin: Add test case for LoadBalancer driver(2)
openstackgerritYanyan Hu proposed stackforge/senlin: Move webhook test case to tests/engine directory
*** jruano has quit IRC06:06
*** Tennyson has quit IRC06:11
*** Tennyson has joined #senlin06:11
*** Zhenqi has quit IRC06:31
openstackgerritYanyan Hu proposed stackforge/senlin: Complete test case of policy base module
Qimingsuccessfully get myself trapped, sigh07:51
Yanyanhuwhat happened07:51
Qiminga policy has 'signa' and 'handlers' section07:52
Qimingthe handler could be a webhook generated out of a CLUSTER_RESIZE action, and the signal can be a ceilometer alarm07:53
Qimingso far so good07:53
Qimingthen this policy cannot be used07:54
Qimingdon't now how to specify its TARGET property07:55
Qimingit doesn't need one, because the triggering signal comes from external systems07:56
Qimingso don't need a ('BEFORE', consts.CLUSTE_RESIZE) for example07:56
Qimingit won't be checked before or after any action execution07:57
Qimingthis is fine too07:57
Qimingalthough we previously defined policy to be something more restrictive, we can revise that definition07:57
Qimingthe real problem is that the handlers may trigger other policies07:58
Qimingit could be a ScalingInPolicy for example07:58
Yanyanhuso the ComplicatedScalingPolicy is not just a 'policy' we imagined before07:59
Qimingno, it is not07:59
Qimingit is gonna break our existing definition07:59
Yanyanhuit's a workflow spec08:00
QimingIt doesn't qualify to be a workflow by itself08:01
Qimingit is just a hook, hooking external signals to senlin actions, via the webhook interface, in a secured way08:01
Qimingthen what are we gonna do with the other policies08:02
Yanyanhubut really not a policy which used to regulated the behavior of cluster action08:02
YanyanhuQiming, not sure whether we should put them together08:04
Yanyanhutreat them as the same thing08:05
QimingI'm still trying to shepherd all things into the same pen08:07
Yanyanhuyes, that's our goal I think, just feel may be there should be some detail classification inside the big pen08:09
*** lixinhui_ has quit IRC08:42
*** mathspanda has quit IRC09:24
*** Tennyson has quit IRC09:32
*** Yanyanhu has quit IRC10:08
*** Qiming has quit IRC10:55
*** Qiming has joined #senlin11:44
*** Qiming has quit IRC12:16
*** Qiming has joined #senlin12:20
*** lkarm has joined #senlin13:09
*** jruano has joined #senlin13:11
openstackgerritQiming Teng proposed stackforge/senlin: Fix environment test case error
jruanoqiming, i spent some time yesterday tracing through the magnum code for bay create13:14
Qimingany findings?13:15
jruanothey use heat to orchestrate their container deployment, and put together a pretty complex template13:15
jruanoi need a little more time today to dig into this a little more. will let you know what i find out13:15
Qiminggreat, thanks13:16
jruanothere is a lot in that template, so im trying to distill some necessary elements to satisfy a poc13:16
Qimingokay, that sounds practical13:17
jruanothe other thing is that there is a process to get kube on coreos13:17
jruanodocker comes built in13:18
jruanowill you require kube for now, or will docker suffice13:18
Qimingmaybe just docker13:18
jruanocool, for now that is the fastest way to a supported runtime13:19
Qimingit seems magnum is using a lot of softwareconfig to deploy kube13:20
jruanoso a lot of elements in the template to coordinate this13:24
jruanoim hoping the docker template is a little simpler13:24
jruanoi can send you a very raw dict output of the template if you want. maybe i can try to pretty it up a little13:24
Qimingthat sounds good, don't waste time pretty it up though, :)13:25
*** jdandrea has joined #senlin13:34
openstackgerritMerged stackforge/senlin: Fix environment test case error
*** lixinhui_ has joined #senlin15:39
*** Qiming has quit IRC15:42
*** lixinhui_ has quit IRC15:59
*** lkarm has left #senlin19:35
*** Qiming has joined #senlin23:32
*** jruano has quit IRC23:34

Generated by 2.14.0 by Marius Gedminas - find it at!