| *** guoshan has joined #senlin | 00:01 | |
| *** guoshan has quit IRC | 00:06 | |
| *** zzxwill has joined #senlin | 00:18 | |
| *** zzxwill has quit IRC | 00:20 | |
| *** zzxwill has joined #senlin | 01:00 | |
| *** yuanying has quit IRC | 01:03 | |
| *** yuanying has joined #senlin | 01:07 | |
| *** guoshan has joined #senlin | 01:14 | |
| *** ruijie has joined #senlin | 01:20 | |
| *** zhurong has joined #senlin | 01:22 | |
| *** yuanying_ has joined #senlin | 01:23 | |
| *** yuanying has quit IRC | 01:26 | |
| *** yanyanhu has joined #senlin | 01:31 | |
| *** zhurong has quit IRC | 01:40 | |
| *** zhurong has joined #senlin | 01:41 | |
| *** elynn has joined #senlin | 01:47 | |
| *** elynn_ has joined #senlin | 01:52 | |
| Qiming | hi, all | 01:52 |
|---|---|---|
| Qiming | hi, guys, http://lists.openstack.org/pipermail/openstack-dev/2016-October/105658.html | 01:52 |
| Qiming | core team, please respond, thanks | 01:52 |
| elynn_ | +1 | 01:53 |
| *** elynn has quit IRC | 01:55 | |
| yanyanhu | +1 from, has replied to the mail | 02:07 |
| yanyanhu | from me | 02:07 |
| yanyanhu | hi, R_lixh, around? | 02:09 |
| yanyanhu | hi, R_lixh, I have written down the design of message type of receiver in the following etherpad. You can refer to it to understand the workflow. | 02:22 |
| yanyanhu | https://etherpad.openstack.org/p/senlin-message-type-receiver | 02:22 |
| Qiming | yanyanhu, why don't we incorporate those texts into the developer docs here? http://git.openstack.org/cgit/openstack/senlin/tree/doc/source/developer/receiver.rst | 02:47 |
| Qiming | I'm really concerned that we are scattering such docs randomly | 02:47 |
| yanyanhu | Qiming, yes, I'm still working on it. Just xinhui asked me a quick introduction for message receiver, I rewrote that etherpad. | 02:48 |
| Qiming | I know you are working on user docs for message type receivers | 02:48 |
| Qiming | this etherpad is more about the design | 02:49 |
| yanyanhu | yes, need to work on developer doc as well | 02:49 |
| Qiming | which I think should be documented | 02:49 |
| yanyanhu | yes | 02:49 |
| Qiming | it would be super helpful for anyone who wants to understand this part of design | 02:50 |
| yanyanhu | will work on developer doc soon to help other developer understand it | 02:50 |
| *** yuanying_ has quit IRC | 02:51 | |
| *** yuanying has joined #senlin | 02:52 | |
| *** guoshan has quit IRC | 02:57 | |
| *** ruijie has quit IRC | 03:00 | |
| *** guoshan has joined #senlin | 03:03 | |
| *** yuanying has quit IRC | 03:15 | |
| *** yuanying has joined #senlin | 03:15 | |
| *** yuanying_ has joined #senlin | 03:17 | |
| *** yuanying has quit IRC | 03:18 | |
| *** yuanying_ has quit IRC | 03:22 | |
| *** ruijie has joined #senlin | 03:25 | |
| openstackgerrit | lvdongbing proposed openstack/senlin: Integrate osprofiler into Senlin https://review.openstack.org/385872 | 03:32 |
| *** shu-mutou-AWAY has quit IRC | 03:41 | |
| *** shu-mutou has joined #senlin | 03:50 | |
| *** yuanying has joined #senlin | 03:52 | |
| *** zzxwill has quit IRC | 04:03 | |
| *** guoshan has quit IRC | 04:05 | |
| *** guoshan has joined #senlin | 04:05 | |
| openstackgerrit | wu.yucheng proposed openstack/senlin-dashboard: Cleanup allowed() for cluster/node action https://review.openstack.org/386327 | 04:09 |
| *** guoshan has quit IRC | 04:10 | |
| *** elynn_ has quit IRC | 04:10 | |
| openstackgerrit | Merged openstack/senlin-dashboard: Cleanup allowed() for cluster/node action https://review.openstack.org/386327 | 04:27 |
| *** xuhaiwei has joined #senlin | 04:42 | |
| openstackgerrit | KATO Tomoyuki proposed openstack/python-senlinclient: Add plug-in summary for osc doc https://review.openstack.org/386334 | 04:47 |
| *** guoshan has joined #senlin | 05:20 | |
| *** guoshan has quit IRC | 05:25 | |
| *** R_lixh has quit IRC | 05:25 | |
| *** elynn_ has joined #senlin | 05:30 | |
| *** elynn__ has joined #senlin | 05:35 | |
| *** elynn_ has quit IRC | 05:35 | |
| *** zzxwill has joined #senlin | 05:36 | |
| ruijie | Hi, Qiming | 05:41 |
| openstackgerrit | Merged openstack/python-senlinclient: Add plug-in summary for osc doc https://review.openstack.org/386334 | 05:50 |
| *** zhurong has quit IRC | 05:51 | |
| *** R_lixh has joined #senlin | 05:59 | |
| *** zzxwill has quit IRC | 06:08 | |
| *** guoshan has joined #senlin | 06:11 | |
| openstackgerrit | lvdongbing proposed openstack/python-senlinclient: Change osprofiler option "--profile" to "--os-profile" https://review.openstack.org/386390 | 06:11 |
| *** lixinhui_ has joined #senlin | 06:20 | |
| *** guoshan_ has joined #senlin | 06:22 | |
| *** guoshan has quit IRC | 06:25 | |
| *** guoshan has joined #senlin | 06:26 | |
| *** guoshan_ has quit IRC | 06:26 | |
| lixinhui_ | yanyanhu, Thanks | 06:31 |
| yanyanhu | lixinhui_, my pleasure. I'm now writing the developer doc of message receiver. Hopefully, the first patch will be proposed this afternoon. | 06:32 |
| yanyanhu | you can read it for more detail information. And plz leave comments if you have any question of confusion about it | 06:32 |
| yanyanhu | s/of/or | 06:33 |
| lixinhui_ | okay | 06:33 |
| Qiming | core team, please check http://lists.openstack.org/pipermail/openstack-dev/2016-October/105658.html | 06:35 |
| Qiming | and leave you vote | 06:35 |
| Qiming | thanks | 06:35 |
| xuhaiwei | Hi Qiming, have some questions about health policy | 06:47 |
| xuhaiwei | I can't find the document about it | 06:47 |
| Qiming | oh, want to write one in Japanese? | 06:47 |
| xuhaiwei | :(, am I going to the wrong place? | 06:48 |
| Qiming | no, the doc for health policy is still to be authored ... | 06:48 |
| Qiming | no doc for developer, no doc for user | 06:49 |
| xuhaiwei | ok, I will explain it to the tacker team, so I need to understand it well | 06:49 |
| Qiming | em ... | 06:49 |
| Qiming | you may want to check with xinhui and ethan | 06:50 |
| Qiming | they are working on a talk on Barcelona ... building auto-healing clusters using senlin | 06:50 |
| xuhaiwei | from the source, I have understood it to some extent | 06:50 |
| Qiming | basically, there will be two questions you will need to answer | 06:51 |
| Qiming | 1) what is HA and why it is needed? | 06:51 |
| Qiming | 2) how could HA be achieved and how to do it with Senlin | 06:51 |
| Qiming | yes, we have a working version of health policy | 06:52 |
| Qiming | but you will have to realize that the road is pretty long because the problem space is too large | 06:52 |
| Qiming | you can communicate that to tacker team | 06:52 |
| Qiming | we are working on a better support to HA even today, to make sure we can cover most typical use cases in real life | 06:54 |
| xuhaiwei | yes, from the source, I see currently if a vm is not in ACTIVE status, senlin will rebuild it or recreate it | 06:54 |
| Qiming | yes you can reference the source code, and let tacker team know where we are today | 06:54 |
| xuhaiwei | I am not understanding well about fencing | 06:55 |
| Qiming | you may also want to let them know that our current design was an open framework | 06:55 |
| Qiming | we considered a lot more use cases before we wrote these code | 06:55 |
| Qiming | for example, auto-healing may not be always valid, we have cluster/node_check, cluster/node_recover APIs added to enable other usage scenarios | 06:56 |
| Qiming | our receivers design allows users to hook any external failure detection mechanism that can trigger a cluster recovery, so on and so forth ... | 06:57 |
| Qiming | fencing is an on-going work | 06:57 |
| Qiming | it is a (necessary) step before doing node recovery | 06:57 |
| Qiming | the reason we need fencing is that failure detection itself cannot be made 100% reliable | 06:58 |
| Qiming | take a VM down event for example | 06:58 |
| Qiming | if the senlin-engine (on management network) cannot reach a VM, it may conclude that VM is dead | 06:59 |
| xuhaiwei | can senlin cooperate with other monitoring tools to check if the vm is healthy or not? | 06:59 |
| Qiming | however, that may and may not be true | 06:59 |
| Qiming | the VM may be still alive on its tenant network | 06:59 |
| Qiming | before going any further on recovering that VM, we will need to **make sure the VM is dead** | 07:00 |
| Qiming | that step is called fencing ... it is about killing a VM by shoot it in the head | 07:00 |
| Qiming | health checking can be completely offloaded to another software/service | 07:00 |
| Qiming | we don't know how users will be doing that | 07:01 |
| Qiming | but we believe they are smart people, and they have tools to monitor their baby VMs | 07:01 |
| Qiming | when they detect such VM failures, they can trigger a senlin receiver to do the recovery | 07:01 |
| Qiming | for those who don't want to bother setting up their own monitoring infra, senlin has a builtin, basic closed-loop | 07:02 |
| Qiming | senlin will detect the VM failure and handle them automatically | 07:02 |
| xuhaiwei | ok, make sense to me, senlin receiver not only manage scale_in/out, but also manage cluster recovery | 07:02 |
| Qiming | yes, it is a socket | 07:03 |
| Qiming | you can use it for whatever legal reasons | 07:03 |
| ruijie | Hi, Qiming. Have a question about the review https://review.openstack.org/#/c/369291/ | 07:05 |
| Qiming | yes? | 07:06 |
| ruijie | Will the action be a new API? | 07:07 |
| ruijie | I mean all the cluster action will be divided to many new APIs | 07:07 |
| Qiming | any user visible API changes should be treated seriously | 07:09 |
| Qiming | this include the introduction of a new URI, new verb, new field into a request, new field into a body, ... | 07:09 |
| Qiming | to avoid bumping the API micro-version frequently, we can group several changes into a single version bump | 07:10 |
| Qiming | for example, last time we bumped api to 1.2 when we introduced three changes to the API | 07:11 |
| Qiming | cluster action changes is a little bit tricky because we are using the same POST URI /clusters/<cid>/actions for action triggering | 07:11 |
| Qiming | that is why I was suggesting you move the new logic out of the current function and annotate it with a new version number | 07:12 |
| openstackgerrit | Merged openstack/python-senlinclient: Change osprofiler option "--profile" to "--os-profile" https://review.openstack.org/386390 | 07:12 |
| Qiming | I'm curious if that will help ensure/improve api versioning | 07:12 |
| *** Qiming is now known as Qiming_afk | 07:13 | |
| ruijie | eh.. that makes sense, I thought we are going to create a new API. Misunderstood the feedback. | 07:13 |
| ruijie | Thanks, Qiming. | 07:15 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: [WIP]Add developer doc for message receiver https://review.openstack.org/386426 | 07:17 |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: [WIP]Add user doc for message receiver https://review.openstack.org/377330 | 07:20 |
| *** Qiming_afk is now known as Qiming | 07:23 | |
| openstackgerrit | Yanyan Hu proposed openstack/senlin: [WIP]Add developer doc for message receiver https://review.openstack.org/386426 | 07:25 |
| openstackgerrit | Shu Muto proposed openstack/senlin-dashboard: Fix typo https://review.openstack.org/386442 | 07:46 |
| xuhaiwei | hi, I am thinking a case that should we add a 'count' option when creating a SCALE_OUT receiver, since sometimes scale out one node is not enough at all | 07:48 |
| xuhaiwei | any ideas about this? | 07:48 |
| Qiming | yes, xuhaiwei, you can do that | 07:50 |
| Qiming | senlin receiver-create | 07:51 |
| Qiming | usage: senlin receiver-create [-t <TYPE>] [-c <CLUSTER>] [-a <ACTION>] | 07:51 |
| Qiming | [-P <KEY1=VALUE1;KEY2=VALUE2...>] | 07:51 |
| Qiming | <NAME> | 07:51 |
| Qiming | you can pass '-P count=3' when creating such a receiver | 07:51 |
| xuhaiwei | oh, I see | 07:52 |
| openstackgerrit | RUIJIE YUAN proposed openstack/senlin: API support node replace https://review.openstack.org/369291 | 07:56 |
| openstackgerrit | Merged openstack/senlin-dashboard: Fix typo https://review.openstack.org/386442 | 07:57 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Add config option 'rpc_use_object' https://review.openstack.org/386446 | 08:05 |
| *** zhurong has joined #senlin | 08:16 | |
| *** shu-mutou is now known as shu-mutou-AWAY | 08:21 | |
| yanyanhu | hi, Qiming, around? | 08:24 |
| yanyanhu | I checked the irc meeting time slot and found utc0600 is available except Tuesday | 08:26 |
| yanyanhu | that will 2pm Beijing time and 3pm tokyo time | 08:26 |
| yanyanhu | will be | 08:26 |
| yanyanhu | hi, xuhaiwei, is that time ok for you? | 08:26 |
| xuhaiwei | yes, it's 11:00 am? | 08:27 |
| yanyanhu | it's 3PM tokyo time | 08:28 |
| yanyanhu | 8AM UTC time | 08:28 |
| yanyanhu | sorry, 6AM UTC time | 08:28 |
| xuhaiwei | i got it | 08:28 |
| xuhaiwei | it's ok | 08:28 |
| yanyanhu | if this time is ok for all people, I will propose to change the weekly irc meeting time | 08:29 |
| xuhaiwei | on what day? | 08:29 |
| yanyanhu | Tuesday is unavalable | 08:29 |
| xuhaiwei | Tuesday? | 08:29 |
| yanyanhu | all other four days are ok | 08:29 |
| yanyanhu | maybe on Wednesday? | 08:29 |
| xuhaiwei | i am ok | 08:29 |
| yanyanhu | Tuesday UTC0600 is now used by doc-install team | 08:30 |
| yanyanhu | I see, will ask other team members | 08:30 |
| yanyanhu | hi, all, is UTC0600 Wednesday a proper time for you guys to join the irc meeting? I have the plan to change our weekly IRC meeting time. If that time slot is ok for all you guys, I will propose the change. | 08:31 |
| yanyanhu | thanks | 08:31 |
| openstackgerrit | RUIJIE YUAN proposed openstack/senlin: API support node replace https://review.openstack.org/369291 | 08:33 |
| Qiming | I'm not sure if holding IRC meetings during work time may cause frequent conflict with company meetings | 08:35 |
| yanyanhu | yes, need to confirm with all guys before making the change | 08:36 |
| Qiming | and UTC0600 would be 14:00 pm in China, 02:00 am in US | 08:37 |
| Qiming | we are totally excluding any guys from other countries to join us, to watch our progress | 08:37 |
| yanyanhu | maybe pick a timeslot in early night of us? | 08:38 |
| Qiming | you can try this: http://xkcd.com/1335/ | 08:38 |
| ruijie | haha, the next pic is fun | 08:40 |
| yanyanhu | :) | 08:40 |
| yanyanhu | yes | 08:40 |
| yanyanhu | Qiming, another possible option is utc0200 which will be 10:am beijing time | 08:42 |
| yanyanhu | for us, it will be 10 pm or 11 pm | 08:43 |
| yanyanhu | still a little late | 08:43 |
| yanyanhu | maybe lets have a further discussion on this topic in next irc meeting | 08:44 |
| yanyanhu | and decide which time most people prefer for the meeting, worktime or night | 08:44 |
| yanyanhu | then we decide to change the irc meeting time or not | 08:45 |
| Qiming | yup, that would be nice | 08:45 |
| openstackgerrit | lvdongbing proposed openstack/senlin: Integrate osprofiler into Senlin https://review.openstack.org/385872 | 08:45 |
| yanyanhu | I see. Will add it to meeting agenda | 08:45 |
| Qiming | okay | 08:54 |
| openstackgerrit | Shu Muto proposed openstack/senlin-dashboard: Cleanup .gitignore and tox.ini https://review.openstack.org/386481 | 08:55 |
| openstackgerrit | Qiming Teng proposed openstack/senlin: Add RPC call for passing object as parameters https://review.openstack.org/386482 | 08:56 |
| *** guoshan has quit IRC | 09:00 | |
| *** guoshan has joined #senlin | 09:01 | |
| *** yanyanhu has quit IRC | 09:08 | |
| *** elynn__ has quit IRC | 09:36 | |
| *** elynn__ has joined #senlin | 09:37 | |
| *** elynn__ has quit IRC | 09:42 | |
| *** elynn__ has joined #senlin | 09:42 | |
| *** lixinhui_ has quit IRC | 09:43 | |
| *** openstackstatus has quit IRC | 09:44 | |
| *** openstack has joined #senlin | 09:49 | |
| *** ruijie has quit IRC | 09:56 | |
| *** zhurong has quit IRC | 10:02 | |
| *** openstackstatus has quit IRC | 10:13 | |
| *** openstack has joined #senlin | 10:17 | |
| *** guoshan_ has joined #senlin | 10:19 | |
| *** guoshan has quit IRC | 10:22 | |
| *** guoshan_ has quit IRC | 10:24 | |
| *** gongysh has quit IRC | 10:56 | |
| *** elynn__ has quit IRC | 11:41 | |
| *** R_lixh has quit IRC | 11:47 | |
| *** openstackstatus has quit IRC | 12:43 | |
| *** openstack has joined #senlin | 12:47 | |
| *** Drago has joined #senlin | 14:35 | |
| *** Drago has joined #senlin | 14:36 | |
| *** Drago has quit IRC | 15:58 | |
| *** Drago has joined #senlin | 16:44 | |
| *** Drago has quit IRC | 17:33 | |
| *** heyongli has quit IRC | 17:40 | |
| *** heyongli has joined #senlin | 17:43 | |
| *** Drago has joined #senlin | 18:45 | |
| *** Drago has quit IRC | 18:45 | |
| *** Drago has joined #senlin | 18:46 | |
| *** Drago has quit IRC | 19:51 | |
| *** Drago has joined #senlin | 19:57 | |
| *** Drago has quit IRC | 21:34 | |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!