| *** yuanying_ has joined #senlin | 00:04 | |
| *** yuanying has quit IRC | 00:05 | |
| *** liyi has joined #senlin | 00:29 | |
| openstackgerrit | liyi proposed openstack/senlin master: Migrate loadbalancer to Octavia https://review.openstack.org/523826 | 00:48 |
|---|---|---|
| *** yuanying_ has quit IRC | 01:29 | |
| *** yuanying has joined #senlin | 01:29 | |
| *** yuanying_ has joined #senlin | 01:33 | |
| *** yuanying has quit IRC | 01:37 | |
| *** ruijie has joined #senlin | 01:53 | |
| ruijie | hi chenyb4 | 01:55 |
| chenyb4 | hi, ruijie | 01:56 |
| ruijie | https://review.openstack.org/#/c/521892/ | 01:56 |
| ruijie | about this patch | 01:57 |
| ruijie | will that cost too much that we stop the listener every time a CLUSTER_RESIZE being processed? | 01:57 |
| chenyb4 | ruijie, listener use oslo.messagine thread start, so , cluster_resize Will there be any problem? | 01:59 |
| chenyb4 | ruijie, cluster_resize use engine-scheduler call thread execute | 01:59 |
| chenyb4 | I didn't find a place for both. | 02:00 |
| ruijie | not really, we will invoke health_manager.disable() before process cluster_action in health policy | 02:01 |
| ruijie | http://git.openstack.org/cgit/openstack/senlin/tree/senlin/policies/health_policy.py#n232 | 02:02 |
| ruijie | looks like the listener itself is also heavy since it is run with a threadpool | 02:03 |
| ruijie | is that possible that we only maintain ONE listener for each health manager to maintain clusters belong to this HM? | 02:05 |
| chenyb4 | ok, I mean. test detach health policy, the messagine will be 5-10s | 02:05 |
| chenyb4 | On listener? rpc listener, get_notification_listener ? share use ? | 02:06 |
| *** ruijie has quit IRC | 02:07 | |
| chenyb4 | all the rpc listener, it's don't require any endpoints. but the event listener must require and execute callback endpoints | 02:09 |
| *** ruijie has joined #senlin | 02:12 | |
| *** liyi has quit IRC | 02:13 | |
| ruijie | network is unstable :( | 02:13 |
| *** liyi has joined #senlin | 02:13 | |
| chenyb4 | ruijie, i think hm.disable maybe don't require | 02:16 |
| chenyb4 | On listener? rpc listener, get_notification_listener ? share use ? | 02:16 |
| chenyb4 | all the rpc listener, it's don't require any endpoints. but the event listener must require and execute callback endpoints | 02:16 |
| chenyb4 | I think, maybe in "NovaNotificationEndpoint" add node filter, if node don't exist, the listener don't any thing | 02:17 |
| ruijie | if we keep listening to the message, we will find some servers may in BAD status, since the servers may in CREATING status | 02:17 |
| ruijie | that's why we want to disable the listener when processing cluster actions | 02:18 |
| ruijie | but as you see, the creating, destroying of listener may cost too much? | 02:19 |
| chenyb4 | if event listener add filter in event trigger after operation, so, the hm.disable i think is not require | 02:19 |
| ruijie | yes, it will, we can add logic to Endpoints | 02:20 |
| ruijie | one listener with mutl-threads can handle this? | 02:22 |
| chenyb4 | Yes. when node delete--> nova shudown.end---> info(check node has exist)---> after operation(when info message trigger) | 02:22 |
| chenyb4 | I don't think so, i think, rpc, event listener should be separated. | 02:23 |
| chenyb4 | event listener only messaging trigger operation or policy detach or disable policy | 02:24 |
| chenyb4 | the other operation did't hm operation | 02:25 |
| ruijie | 1. | 02:25 |
| chenyb4 | We can detect the accident node at the time of the event, determine whether the node performs the recovery operation, which is controllable | 02:27 |
| ruijie | http://paste.openstack.org/show/627780/ | 02:28 |
| *** shu-mutou-AWAY is now known as shu-mutou | 02:30 | |
| chenyb4 | Example: Special status filter, Delete node has exist check. | 02:30 |
| chenyb4 | yes, i think so. and add detach policy | 02:30 |
| *** liyi_ has joined #senlin | 02:30 | |
| chenyb4 | detach policy stop listener | 02:30 |
| ruijie | 起停listener的开销很大 | 02:31 |
| chenyb4 | 如果我们对所有事件进行监听(一个监听),根据不同的cluster_id进行消费,这样看起来很合理,但是不好控制 | 02:33 |
| *** liyi has quit IRC | 02:34 | |
| *** ruijie has quit IRC | 02:35 | |
| *** liyi_ has quit IRC | 03:01 | |
| *** liyi has joined #senlin | 03:02 | |
| *** liyi_ has joined #senlin | 03:23 | |
| *** liyi_ has quit IRC | 03:24 | |
| *** liyi has quit IRC | 03:24 | |
| *** liyi has joined #senlin | 03:24 | |
| *** liyi has quit IRC | 03:32 | |
| *** liyi has joined #senlin | 04:14 | |
| liyi | https://review.openstack.org/#/c/523826/ your suggestion is welcome | 04:15 |
| *** AlexeyAbashkin has joined #senlin | 04:15 | |
| liyi | pls review this path, thanks!@alll | 04:15 |
| liyi | pls review this patch, thanks! @alll | 04:16 |
| *** AlexeyAbashkin has quit IRC | 04:20 | |
| *** Jeffrey4l has quit IRC | 04:42 | |
| *** Jeffrey4l has joined #senlin | 05:04 | |
| *** ruijie has joined #senlin | 06:08 | |
| openstackgerrit | liyi proposed openstack/senlin master: Migrate loadbalancer to Octavia https://review.openstack.org/523826 | 06:20 |
| *** XueFeng has quit IRC | 06:32 | |
| *** XueFeng has joined #senlin | 06:33 | |
| *** XueFeng has quit IRC | 06:35 | |
| *** XueFeng has joined #senlin | 06:36 | |
| *** liyi has quit IRC | 06:57 | |
| *** liyi has joined #senlin | 07:00 | |
| *** liyi has quit IRC | 07:00 | |
| *** liyi has joined #senlin | 07:02 | |
| *** AlexeyAbashkin has joined #senlin | 08:06 | |
| openstackgerrit | Merged openstack/senlin master: Add cluster action to profile https://review.openstack.org/523054 | 08:16 |
| *** AlexeyAbashkin has quit IRC | 08:41 | |
| *** AlexeyAbashkin has joined #senlin | 08:42 | |
| *** liyi_ has joined #senlin | 09:19 | |
| *** liyi has quit IRC | 09:22 | |
| *** ruijie has quit IRC | 09:31 | |
| *** liyi_ has quit IRC | 10:05 | |
| *** zhenguo has quit IRC | 10:09 | |
| *** zhenguo has joined #senlin | 10:15 | |
| *** shu-mutou is now known as shu-mutou-AWAY | 10:55 | |
| *** jmlowe has quit IRC | 11:01 | |
| *** liyi has joined #senlin | 11:22 | |
| openstackgerrit | liyi proposed openstack/senlin master: Migrate loadbalancer to Octavia https://review.openstack.org/523826 | 11:37 |
| *** liyi has quit IRC | 12:09 | |
| *** chenybd_ has joined #senlin | 12:51 | |
| *** jmlowe has joined #senlin | 13:19 | |
| *** chenybd_ has quit IRC | 14:15 | |
| *** chenybd_ has joined #senlin | 14:15 | |
| *** chenybd_ has quit IRC | 14:15 | |
| *** chenybd_ has joined #senlin | 14:16 | |
| *** jmlowe has quit IRC | 14:24 | |
| *** chenybd__ has joined #senlin | 14:26 | |
| *** chenybd_ has quit IRC | 14:29 | |
| *** chenybd_ has joined #senlin | 14:39 | |
| *** chenybd__ has quit IRC | 14:41 | |
| *** chenybd_ has quit IRC | 14:53 | |
| *** chenybd_ has joined #senlin | 14:54 | |
| *** chenybd_ has quit IRC | 14:58 | |
| *** jmlowe has joined #senlin | 15:19 | |
| -openstackstatus- NOTICE: if you receieved a result of "RETRY_LIMIT" after 14:15 UTC, it was likely due to an error since corrected. please "recheck" | 15:39 | |
| *** AlexeyAbashkin has quit IRC | 17:01 | |
| *** jmlowe has quit IRC | 17:06 | |
| *** jmlowe has joined #senlin | 17:31 | |
| *** openstackgerrit has quit IRC | 18:48 | |
| *** AlexeyAbashkin has joined #senlin | 19:52 | |
| *** AlexeyAbashkin has quit IRC | 19:59 | |
| *** AlexeyAbashkin has joined #senlin | 20:12 | |
| *** AlexeyAbashkin has quit IRC | 20:17 | |
| *** jmlowe has quit IRC | 22:58 | |
| *** jmlowe has joined #senlin | 23:09 | |
| *** AlexeyAbashkin has joined #senlin | 23:12 | |
| *** AlexeyAbashkin has quit IRC | 23:16 | |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!