*** liyi has joined #senlin | 00:19 | |
*** XueFeng has quit IRC | 00:20 | |
*** XueFeng has joined #senlin | 00:20 | |
*** liyi has quit IRC | 00:23 | |
*** XueFeng has quit IRC | 00:33 | |
*** XueFeng has joined #senlin | 00:33 | |
*** liyi has joined #senlin | 00:48 | |
openstackgerrit | Merged openstack/senlin master: Fix nova handle_rebuild return error https://review.openstack.org/518651 | 01:34 |
---|---|---|
liyi | hi all, I have a question about lb ceration. https://github.com/openstack/senlin/blob/master/senlin/drivers/openstack/lbaas.py#L66, The code shows that when we create a lb, we need to check provisioning_status and operating_status, but when lb is created at the first time, the operating_status must be offline, because no member in lb. | 01:58 |
ruijie_ | the response sample on LBaas's api documentation shows that the "operating_status == ONLINE" ? | 02:09 |
*** shu-mutou-AWAY has quit IRC | 02:09 | |
*** liyi has quit IRC | 02:11 | |
*** liyi has joined #senlin | 02:11 | |
*** liyi_ has joined #senlin | 02:13 | |
*** liyi has quit IRC | 02:13 | |
liyi_ | ONLINE | 02:14 |
liyi_ | Entity is operating normally | 02:14 |
liyi_ | All pool members are healthy | 02:14 |
liyi_ | ruijie, if members are healthy, operating_status is ONLINE | 02:14 |
liyi_ | but the lb initiate with no any member, so the operating_status must be OFFLINE, | 02:15 |
liyi_ | I don't know right answer | 02:16 |
liyi_ | https://developer.openstack.org/api-ref/load-balancer/v2/index.html#operating-status-codes | 02:17 |
liyi_ | Is it necessary to check operating_status is ONLINE? | 02:20 |
*** XueFeng has quit IRC | 02:35 | |
*** XueFeng has joined #senlin | 02:35 | |
ruijie_ | not sure about that :< | 02:36 |
openstackgerrit | Merged openstack/senlin master: Remove redundant character. https://review.openstack.org/523345 | 02:38 |
chenyb4 | Hi, all! I have a question. when engine service stop or restart don't stop all thread, only stop one thread, the other stop appear "NoneType" error. look this bug https://bugs.launchpad.net/senlin/+bug/1735064 | 04:29 |
openstack | Launchpad bug 1735064 in senlin "Engine service stop or restart 'NoneType' object has no attribute 'stop'" [Undecided,New] - Assigned to chenyb4 (chenyb4) | 04:29 |
*** shu-mutou has joined #senlin | 04:29 | |
chenyb4 | So, the engine-service stop, It doesn't seem to be thinking about "num_workers" | 04:30 |
openstackgerrit | chenyb4 proposed openstack/senlin master: Fix node check no server found record error https://review.openstack.org/523692 | 05:35 |
*** liyi has joined #senlin | 05:56 | |
*** liyi_ has quit IRC | 05:57 | |
chenyb4 | Hi, all! I have a question. when engine service stop or restart don't stop all thread, only stop one thread, the other stop appear "NoneType" error. look this bug https://bugs.launchpad.net/senlin/+bug/1735064 | 06:06 |
openstack | Launchpad bug 1735064 in senlin "Engine service stop or restart 'NoneType' object has no attribute 'stop'" [Undecided,New] - Assigned to chenyb4 (chenyb4) | 06:06 |
chenyb4 | So, the engine-service stop, It doesn't seem to be thinking about "num_workers" | 06:07 |
Qiming | what do you mean, chenyb4 ? | 06:09 |
Qiming | there are two layers of "workers" | 06:09 |
Qiming | the configuration option "workers" actually means the number of engine-services you are running | 06:09 |
Qiming | inside each engine-service instances there are many threads | 06:10 |
chenyb4 | Qiming, https://github.com/openstack/senlin/blob/master/senlin/engine/service.py#L166 In this, the engine stop only stop one | 06:11 |
Qiming | yes | 06:12 |
Qiming | each engine-service has an associated dispatcher | 06:12 |
chenyb4 | So, I think must stop all | 06:12 |
Qiming | that engine-service is only responsible to kill its own dispatcher | 06:12 |
Qiming | having multiple dispatcher is to maintain the availability of the senlin engine service | 06:13 |
chenyb4 | But mutiple dispatcher start after execute stop, will appear "ERROR senlin-engine AttributeError: 'EngineService' object has no attribute 'dispatcher'" | 06:14 |
Qiming | which line is throwing that exceptoin? | 06:14 |
chenyb4 | https://github.com/openstack/senlin/blob/master/senlin/engine/service.py#L166 in this call | 06:15 |
Qiming | the engine-service instance is not yet stopped, the current logic seems correct | 06:15 |
Qiming | you mean that call or that line? | 06:15 |
chenyb4 | yes | 06:16 |
Qiming | I'm asking question | 06:16 |
Qiming | do you mean that "call" ? or that line? | 06:16 |
Qiming | self.dispatcher is initialized at line 116 | 06:16 |
Qiming | it was never changed later | 06:16 |
Qiming | how could it become None? | 06:17 |
chenyb4 | when i update to this, the error, don't appear. http://paste.ubuntu.com/26069902/ | 06:17 |
Qiming | it has nothing to do how it seems working | 06:17 |
Qiming | we don't make decisions based on testing | 06:18 |
Qiming | we have to ensure we are doing the right thing | 06:18 |
chenyb4 | in this, I record log it has nothing to do how it seems working | 06:18 |
Qiming | it is dangerous if an engine-service doesn't have a dispatcher started | 06:18 |
chenyb4 | sorry | 06:18 |
chenyb4 | in this, I record log https://bugs.launchpad.net/senlin/+bug/1735064 | 06:19 |
openstack | Launchpad bug 1735064 in senlin "Engine service stop or restart 'NoneType' object has no attribute 'stop'" [Undecided,New] - Assigned to chenyb4 (chenyb4) | 06:19 |
chenyb4 | start is not problem, only stop appear error | 06:20 |
Qiming | sigh ... | 06:22 |
Qiming | 看样子你听不懂 ... | 06:22 |
Qiming | 问题不在于你修改后的代码是不是测试没有问题 | 06:23 |
Qiming | 问题在于背后的逻辑是不是对的 | 06:23 |
Qiming | 你这样修改之后,默认就认为engine启动之后,dispatcher没有创建也是没有关系的,stop可以正确结束了 | 06:23 |
Qiming | 这是错的 | 06:23 |
Qiming | dispatcher必须启动,不然engine没有意义 | 06:24 |
Qiming | 还可能带来其它不可预料的问题 | 06:24 |
chenyb4 | 我明白你的意思了,在服务停止时,每一个workers都会执行stop操作,那么就不会存在None type | 06:24 |
chenyb4 | 是这样的吗? | 06:24 |
Qiming | worker这个名字不豪 | 06:24 |
Qiming | 不好 | 06:24 |
Qiming | 其实是engines | 06:24 |
Qiming | 每个engine自己启动的时候,设置自己的dispatcher,必须成功,不然engine没用了 | 06:24 |
Qiming | 结束的时候,当然dispatcher不应该为空 | 06:25 |
Qiming | 如果self.dispatcher为空,证明engine出大问题了 | 06:25 |
Qiming | 懂了么? | 06:25 |
chenyb4 | 嗯,明白了. | 06:25 |
Qiming | 所以问题的本质是,engine是怎么停掉的 | 06:25 |
Qiming | 这个调用流程是不是出问题了 | 06:26 |
Qiming | 怎么会启动的时候初始化了self.dispatcher,stop的时候变成NoneType了 | 06:26 |
Qiming | 例如,所有的后台service都不应该用control-C来停止 | 06:28 |
chenyb4 | 服务启动后,我执行ps -ef|grep senlin-engine, 多个engines是存在的, 当我执行stop操作的时候,虽然日志出现了错误,但是senlin-engine进程是结束了的 | 06:28 |
Qiming | Control-C发送的是SIGINT,是对整个引擎运行状态的破坏性操作 | 06:29 |
Qiming | 停止服务的正确信号应该是SIGTERM | 06:29 |
chenyb4 | Qiming, 我明白你的意思, 我使用的是systemctl 管理 | 06:29 |
Qiming | 相关的控制逻辑比较复杂 | 06:29 |
Qiming | 在 senlin/api/common/wsgi.py 里面 | 06:30 |
Qiming | 所以先不要这样去修 | 06:30 |
chenyb4 | 好的. | 06:30 |
Qiming | 从头看一下杀死service-engine的流程是不是哪里有大问题 | 06:31 |
chenyb4 | 好 | 06:31 |
Qiming | 最近一直很忙,没有参与这些讨论,很抱歉 | 06:31 |
*** liyi has quit IRC | 07:16 | |
*** liyi has joined #senlin | 07:17 | |
*** liyi has quit IRC | 07:17 | |
*** liyi has joined #senlin | 07:18 | |
*** AlexeyAbashkin has joined #senlin | 07:47 | |
*** XueFeng has quit IRC | 07:59 | |
*** XueFeng has joined #senlin | 07:59 | |
*** XueFeng has quit IRC | 08:37 | |
*** XueFeng has joined #senlin | 08:38 | |
*** XueFeng has quit IRC | 08:39 | |
*** XueFeng has joined #senlin | 08:40 | |
openstackgerrit | OpenStack Proposal Bot proposed openstack/senlin master: Updated from global requirements https://review.openstack.org/523797 | 09:22 |
openstackgerrit | liyi proposed openstack/senlin master: Merge loadbalancer to Octavia https://review.openstack.org/523826 | 09:48 |
openstackgerrit | liyi proposed openstack/senlin master: Migrate loadbalancer to Octavia https://review.openstack.org/523826 | 09:50 |
*** liyi has quit IRC | 09:53 | |
*** shu-mutou is now known as shu-mutou-AWAY | 10:21 | |
*** AlexeyAbashkin has quit IRC | 10:28 | |
*** liyi has joined #senlin | 11:16 | |
*** haint_ has joined #senlin | 11:16 | |
*** AlexeyAbashkin has joined #senlin | 11:19 | |
*** liyi has quit IRC | 11:20 | |
*** haint has quit IRC | 11:21 | |
*** AlexeyAbashkin has quit IRC | 11:23 | |
*** AlexeyAbashkin has joined #senlin | 11:27 | |
*** ruijie_ has quit IRC | 12:31 | |
*** liyi has joined #senlin | 13:14 | |
*** chenybd_ has joined #senlin | 13:15 | |
*** liyi has quit IRC | 13:18 | |
*** zhurong has joined #senlin | 13:38 | |
*** openstackstatus has quit IRC | 13:43 | |
*** openstack has joined #senlin | 13:47 | |
*** ChanServ sets mode: +o openstack | 13:47 | |
*** zhurong has quit IRC | 13:56 | |
*** zhurong has joined #senlin | 13:58 | |
*** zhurong has quit IRC | 14:46 | |
*** zhurong has joined #senlin | 14:46 | |
*** openstack has quit IRC | 14:46 | |
*** openstack has joined #senlin | 14:53 | |
*** ChanServ sets mode: +o openstack | 14:53 | |
*** liyi has joined #senlin | 14:55 | |
*** liyi has quit IRC | 14:59 | |
*** chenybd_ has quit IRC | 15:49 | |
*** chenybd_ has joined #senlin | 15:49 | |
*** chenybd_ has quit IRC | 15:54 | |
*** AlexeyAbashkin has quit IRC | 16:23 | |
*** liyi has joined #senlin | 16:24 | |
*** liyi has quit IRC | 16:29 | |
*** liyi has joined #senlin | 17:44 | |
*** liyi has quit IRC | 17:49 | |
openstackgerrit | Duc Truong proposed openstack/senlin master: Add lifecycle hook spec https://review.openstack.org/523965 | 19:10 |
*** AlexeyAbashkin has joined #senlin | 20:14 | |
*** AlexeyAbashkin has quit IRC | 20:18 | |
*** liyi has joined #senlin | 20:32 | |
*** liyi has quit IRC | 20:37 | |
*** jmlowe has quit IRC | 22:16 | |
*** liyi has joined #senlin | 22:22 | |
*** liyi has quit IRC | 22:26 | |
*** jmlowe has joined #senlin | 22:31 | |
*** liyi has joined #senlin | 23:11 | |
*** jmlowe has quit IRC | 23:13 | |
*** AlexeyAbashkin has joined #senlin | 23:14 | |
*** liyi has quit IRC | 23:16 | |
*** AlexeyAbashkin has quit IRC | 23:19 | |
*** jmlowe has joined #senlin | 23:21 |
Generated by irclog2html.py 2.15.3 by Marius Gedminas - find it at mg.pov.lt!