| *** Qiming has quit IRC | 00:11 | |
| *** zzxwill has joined #senlin | 00:15 | |
| openstackgerrit | Merged openstack/senlin: Refactor health manager for future work https://review.openstack.org/284193 | 00:20 |
|---|---|---|
| *** Qiming has joined #senlin | 01:06 | |
| openstackgerrit | Merged openstack/senlin: Refactor scaling policy implementation https://review.openstack.org/284011 | 01:13 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Fix target capacity calculation https://review.openstack.org/284495 | 01:38 |
| *** Liuqing has joined #senlin | 01:51 | |
| *** Yanyanhu has joined #senlin | 01:59 | |
| openstackgerrit | Merged openstack/senlin: Revised TODO list https://review.openstack.org/283915 | 02:19 |
| openstackgerrit | Merged openstack/senlin: Fix target capacity calculation https://review.openstack.org/284495 | 02:54 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Documentation for deletion policy https://review.openstack.org/283558 | 02:57 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Documentation for scaling policy v1.0 https://review.openstack.org/284504 | 02:57 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: Minor revise scaling_policy functional test https://review.openstack.org/284509 | 03:15 |
| Yanyanhu | hi, Qiming, did you meet this error when listing cluster? | 03:15 |
| Yanyanhu | 'HttpException' object has no attribute 'status_code' | 03:15 |
| Yanyanhu | all other list operations work correctly | 03:16 |
| Yanyanhu | just cluster-list will give this error | 03:16 |
| Yanyanhu | using latest senlin branch | 03:16 |
| *** yuanying has quit IRC | 03:16 | |
| Yanyanhu | looks like I need to update my senlinclient | 03:19 |
| Yanyanhu | show_nested has been removed | 03:19 |
| Yanyanhu | but this is still an unexpected error msg returned to user | 03:19 |
| Qiming | Yanyanhu, no I don't see that error | 03:28 |
| *** Liuqing has quit IRC | 03:37 | |
| *** zzxwill has quit IRC | 03:37 | |
| *** Liuqing has joined #senlin | 03:38 | |
| *** yuanying has joined #senlin | 04:07 | |
| *** Liuqing has quit IRC | 04:21 | |
| *** Liuqing has joined #senlin | 04:22 | |
| *** lixinhui has joined #senlin | 04:27 | |
| *** lixinhui has quit IRC | 04:34 | |
| *** zzxwill has joined #senlin | 04:58 | |
| *** Liuqing has quit IRC | 05:08 | |
| *** Liuqing has joined #senlin | 05:08 | |
| *** sridhar_ram has joined #senlin | 05:12 | |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Fix scaling functional test https://review.openstack.org/284540 | 05:13 |
| openstackgerrit | Merged openstack/senlin-dashboard: Add proper created_at in receiver detail page https://review.openstack.org/284021 | 05:15 |
| Yanyanhu | hi, Qiming, I have proposed a patch for functional test :) | 05:15 |
| Yanyanhu | https://review.openstack.org/284509 | 05:15 |
| Yanyanhu | but seems the gate has some problems to setup the test env | 05:16 |
| Yanyanhu | so the devstack installation always failed for this error: ERROR! Specified hosts options do not match any hosts | 05:16 |
| Qiming | okay | 05:17 |
| Qiming | that is the gate problem | 05:17 |
| Yanyanhu | yes, will wait for a while | 05:17 |
| *** yuanying has quit IRC | 05:47 | |
| *** sridhar_ram has quit IRC | 06:17 | |
| *** Liuqing has quit IRC | 06:53 | |
| *** yuanying has joined #senlin | 06:54 | |
| *** yuanying has quit IRC | 08:01 | |
| *** yuanying has joined #senlin | 08:11 | |
| *** yuanying has quit IRC | 08:16 | |
| *** yuanying has joined #senlin | 08:18 | |
| *** idonotknow_ has joined #senlin | 08:21 | |
| idonotknow_ | Yanyanhu,I am an intern in vmware of R_lixh. I am confused about ceilometer alarm,could you help me? | 08:24 |
| Yanyanhu | idonotknow_, sure, what issue did you meet | 08:25 |
| *** yuanying has quit IRC | 08:26 | |
| idonotknow_ | Yanyanhu, I want to use ceilometer to monitor all of members's status in a cluster, that means active or inactive, but this can not be described as a threshold. | 08:28 |
| Yanyanhu | you mean the status itself? | 08:29 |
| Yanyanhu | in this case, I think you can define a metric which has two states, 0 and 1 | 08:29 |
| Yanyanhu | 0 means active, while 1 means inactive | 08:29 |
| Yanyanhu | then you can use 1 as the threshold I guess | 08:29 |
| idonotknow_ | So I should change pipeline.yaml or where can I define this? | 08:30 |
| Yanyanhu | yes, I think so | 08:30 |
| Yanyanhu | you need to change the pipeline to add this metric you are interested in | 08:30 |
| Yanyanhu | you also need to fill in the correct value of metric when generating a sample | 08:31 |
| *** yuanying has joined #senlin | 08:32 | |
| idonotknow_ | I want to meter network.services.lb.member which is not included in meter-list,deos that matter? | 08:32 |
| Yanyanhu | for those meters that are not supported by ceilometer, you may need to define it by your self | 08:33 |
| Yanyanhu | I remember ceilometer allows users to define their own metrics | 08:33 |
| Yanyanhu | but that also means users have to generate sample by themselves | 08:34 |
| Yanyanhu | for those built-in metrics, you can just add related configuration in pipeline.yaml to enable them | 08:34 |
| idonotknow_ | It exists in the source code through yield sample.Sample(XXX) | 08:35 |
| Yanyanhu | you mean ceilometer has provided the routine to collect this kind of meter and generated sample for it? | 08:35 |
| Yanyanhu | an easy way to verify this is to list samples with specific meter type | 08:37 |
| idonotknow_ | yes,I can get the meter without change any source code or conf file. cause there is * in pipeline.yaml which can produce all the existed meter I guess | 08:37 |
| Yanyanhu | yes, you just need to change it if you need extra calculation based on built-in meter | 08:37 |
| Yanyanhu | like cpu_utils which is based on cpu | 08:38 |
| Yanyanhu | if so, that's much easier | 08:38 |
| Yanyanhu | you can just define an alarm with that meter type | 08:38 |
| Qiming | I don't think a computation is necessary at all | 08:38 |
| *** yuanying has quit IRC | 08:38 | |
| Yanyanhu | ceilometer will help you to evaluate the samples and trigger alarm when conditions are satisfied | 08:39 |
| Yanyanhu | Qiming, you mean the cpu_utils? | 08:39 |
| idonotknow_ | But status can not be described as a threshold | 08:40 |
| Qiming | ceilometer alarm-create --name lb_alarm_1 --enabled True --alarm-action http://somewhre --period 60 --evaluation-periods 1 -m network.services.lb.member --statistics count --comparison-overator lt --threshold 1 --matching-metadata some_key=somevalue | 08:41 |
| Qiming | why not? | 08:41 |
| Qiming | you can use 'count' | 08:41 |
| Qiming | it is not a metric value for you to do maths | 08:41 |
| *** yuanying has joined #senlin | 08:41 | |
| Yanyanhu | count means once a sample is collected, alarm will be triggered | 08:41 |
| Qiming | the operator is 'lt' | 08:42 |
| Yanyanhu | this is suitable for inactive | 08:42 |
| Qiming | it means if the sample is NOT collected | 08:42 |
| Qiming | this is how you do heatbeat kind of node activity detection, just like this: http://git.openstack.org/cgit/openstack/heat-templates/tree/cfn/F17/WordPress_Single_Instance_With_IHA.template#n105 | 08:43 |
| idonotknow_ | Thank you very much | 08:43 |
| Yanyanhu | so no obvious obstacle there I think :) | 08:43 |
| Qiming | just count if you can get the metric or not | 08:43 |
| Qiming | the tricky part is about --matching-metadata | 08:44 |
| Qiming | you got to find out how to filter out the correct sample | 08:44 |
| idonotknow_ | Maybe I get it, I will give it a try. Thank you~ | 08:45 |
| *** yuanying has quit IRC | 08:47 | |
| Qiming | pay attention to the matching_metadata part, I always get that incorrectly specified | 09:05 |
| idonotknow_ | OK,I will be careful about this | 09:05 |
| Yanyanhu | there is no document to clearly describe how to write the match condition correctly :) | 09:06 |
| Yanyanhu | guess so | 09:07 |
| Yanyanhu | hi, Qiming, do we have plan to support get_file in senlinclient? | 09:16 |
| Yanyanhu | I remeber ethan abandoned his patch about that support finally | 09:16 |
| Qiming | use case? | 09:18 |
| Yanyanhu | currently, if we want to embed a big script into nova server's user_data, we can only define it as a string? | 09:18 |
| Qiming | we may have to | 09:32 |
| *** zzxwill has quit IRC | 09:32 | |
| Yanyanhu | that will make the user_data section of nova server profile very difficult to read | 09:33 |
| Yanyanhu | since there could be lots of line in some scripts | 09:33 |
| Yanyanhu | like the bigdata cluster they are using now | 09:33 |
| Yanyanhu | and lots of nested strings | 09:34 |
| Qiming | yep, I know that | 09:35 |
| Yanyanhu | or is there any built-in support in yaml loader can address this issue? | 09:36 |
| Qiming | yes, there is | 09:37 |
| Qiming | this is something I tried a long time ago | 09:39 |
| Qiming | env: &ENVIRON1 | 09:39 |
| Qiming | parameters: | 09:39 |
| Qiming | len: 16 | 09:39 |
| Qiming | profiles: | 09:39 |
| Qiming | os.docker: some.path.docker.driver | 09:39 |
| Qiming | # preserve newline | 09:39 |
| Qiming | logo: | | 09:39 |
| Qiming | .-- .-- . . . . . . | 09:39 |
| Qiming | |_ |-- |\| | | |\| | 09:39 |
| Qiming | __| |__ | | L__ | . . | 09:39 |
| Qiming | # don't preserve newline | 09:39 |
| Qiming | description: > | 09:39 |
| Qiming | Some description about | 09:39 |
| Qiming | the usage of the stack | 09:39 |
| Qiming | to be created | 09:39 |
| Qiming | for example | 09:39 |
| Qiming | this line | 09:39 |
| Qiming | back to normal | 09:39 |
| Qiming | type: os.heat.stack | 09:39 |
| Qiming | spec: | 09:39 |
| Qiming | # template: !include subdir/subdir/my_stack.yaml | 09:39 |
| Qiming | environment: *ENVIRON1 | 09:39 |
| Qiming | ~ | 09:39 |
| Qiming | you can try the way I was including a environment | 09:40 |
| *** zzxwill has joined #senlin | 09:40 | |
| Yanyanhu | the include keyword? | 09:40 |
| Qiming | no, include is a variant of get_file | 09:40 |
| Qiming | it is an extension of yaml loader | 09:40 |
| Qiming | the environment part | 09:40 |
| Qiming | that is the way you reference something from outside the current structure | 09:40 |
| Yanyanhu | ok, I see | 09:41 |
| Yanyanhu | will have a try. | 09:41 |
| Yanyanhu | hi, Qiming where is env: &ENVIRON1 defined? | 09:48 |
| Yanyanhu | the same file as os.heat.stack spec? | 09:48 |
| Yanyanhu | oh, I see | 09:48 |
| Yanyanhu | still can't load it correctly | 09:50 |
| Yanyanhu | will keep trying to tomorrow. Using string works fine. | 09:53 |
| Qiming | &IDENTIFYER | 09:54 |
| Qiming | should be used in pair with *IDENTIFIER | 09:54 |
| Yanyanhu | yes. And they should defined in the same file? | 09:54 |
| Qiming | 'env: &ENVIRON1' this defined an ankor | 09:55 |
| Qiming | 'environment: *ENVIRON1' this defines a link | 09:55 |
| Yanyanhu | oh, so 'environment' is also a keywork | 09:55 |
| Yanyanhu | so I should define something like: user_data: {enviornment: &ENVNAME} | 09:56 |
| Qiming | no | 09:56 |
| Qiming | user_data: *USER_DATA_SCRIPT | 09:56 |
| Qiming | then you have a key named whatever | 09:56 |
| Qiming | my_script: &USER_DATA_SCRIPT | 09:56 |
| Qiming | my_script is moved to the top level | 09:57 |
| Yanyanhu | ok, let me have a try again. | 09:57 |
| Qiming | this is my preferred way of embedding scripts | 10:02 |
| Yanyanhu | the keyname of 'my_script' always be recognized as a spec item | 10:02 |
| Qiming | the logic is I really hate having separate files from templates | 10:02 |
| Qiming | that is something we can fix from client side | 10:03 |
| Yanyanhu | sure | 10:03 |
| Yanyanhu | all construction should be done before template reaches service side | 10:03 |
| Qiming | we can ignore all properties other than 'type', 'version' and 'properties' after parsing | 10:03 |
| Qiming | yes, keeping things in a single file helps manage a complete set of things | 10:04 |
| Qiming | or else when you change your script separately, we don't know if that is a new profile or not | 10:04 |
| Yanyanhu | anyway, directly using string works fine. If user want to make the template more clear, they can use this way to define 'variable' | 10:04 |
| Yanyanhu | for each single script | 10:04 |
| Qiming | yes | 10:06 |
| *** Yanyanhu has quit IRC | 10:15 | |
| *** zzxwill has quit IRC | 10:16 | |
| *** Qiming has quit IRC | 10:17 | |
| *** zzxwill has joined #senlin | 10:21 | |
| zzxwill | In the section "Manual Installation" of http://docs.openstack.org/developer/senlin/install.html, there should be no <DB PASSWORD> or <SENLIN PASSWORD> in Step 4 without Step 5 being executed. | 10:27 |
| *** shu-mutou is now known as shu-mutou-AWAY | 10:29 | |
| *** Qiming has joined #senlin | 11:04 | |
| *** Liuqing has joined #senlin | 12:03 | |
| *** Liuqing has quit IRC | 12:16 | |
| *** idonotknow_ has quit IRC | 12:41 | |
| openstackgerrit | Merged openstack/python-senlinclient: Add OpenstackClient plugin for cluster policy create https://review.openstack.org/282818 | 12:52 |
| openstackgerrit | Merged openstack/senlin: Add some backlog of release notes https://review.openstack.org/283909 | 12:59 |
| openstackgerrit | Merged openstack/python-senlinclient: Add OpenstackClient plugin for cluster policy update https://review.openstack.org/282819 | 13:03 |
| openstackgerrit | Merged openstack/python-senlinclient: Add OpenstackClient plugin for cluster policy delete https://review.openstack.org/282820 | 13:06 |
| openstackgerrit | Cindia-blue proposed openstack/senlin: Add Registry Table for Health Management https://review.openstack.org/284707 | 13:13 |
| zzxwill | In the section "Manual Installation" of http://docs.openstack.org/developer/senlin/install.html, there should be no <DB PASSWORD> or <SENLIN PASSWORD> in Step 4 without Step 5 being executed. | 13:13 |
| Qiming | zzxwill, thanks, can you help propose a fix? | 13:14 |
| zzxwill | Sure. | 13:14 |
| openstackgerrit | Merged openstack/python-senlinclient: OSC plugin for openstack cluster list https://review.openstack.org/284104 | 13:15 |
| Qiming | great! | 13:15 |
| openstackgerrit | Cindia-blue proposed openstack/senlin: Add Registry Table for Health Management https://review.openstack.org/284707 | 13:15 |
| openstackgerrit | Di XiaoLi proposed openstack/python-senlinclient: OSC plugin for openstack cluster policy type list https://review.openstack.org/282865 | 13:59 |
| openstackgerrit | Di XiaoLi proposed openstack/python-senlinclient: Add OSC plugin for cluster policy type show https://review.openstack.org/282866 | 14:04 |
| openstackgerrit | Cindia-blue proposed openstack/senlin: Add Registry Table for Health Management https://review.openstack.org/284707 | 14:28 |
| openstackgerrit | Cindia-blue proposed openstack/senlin: Add Registry Table for Health Management https://review.openstack.org/284707 | 14:33 |
| openstackgerrit | Cindia-blue proposed openstack/senlin: Add Registry Table for Health Management https://review.openstack.org/284707 | 14:42 |
| openstackgerrit | Cindia-blue proposed openstack/senlin: Add Registry Table for Health Management https://review.openstack.org/284707 | 14:50 |
| *** sridhar_ram has joined #senlin | 15:02 | |
| *** Qiming has quit IRC | 15:28 | |
| *** zzxwill has quit IRC | 15:57 | |
| *** idonotknow_ has joined #senlin | 16:33 | |
| *** sridhar_ram has quit IRC | 16:43 | |
| *** sridhar_ram has joined #senlin | 17:55 | |
| *** sridhar_ram has quit IRC | 18:02 | |
| *** sridhar_ram has joined #senlin | 18:02 | |
| *** sridhar_ram has quit IRC | 18:57 | |
| *** sridhar_ram has joined #senlin | 19:01 | |
| *** idonotknow_ has quit IRC | 19:26 | |
| *** Qiming has joined #senlin | 23:26 | |
| *** yuanying has joined #senlin | 23:33 | |
| *** yuanying has quit IRC | 23:34 | |
| *** yuanying has joined #senlin | 23:34 | |
| *** sridhar_ram has quit IRC | 23:44 | |
| *** Qiming has quit IRC | 23:56 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!