14:00:52 <joehuang> #startmeeting tricircle
14:00:52 <openstack> Meeting started Wed Apr 19 14:00:52 2017 UTC and is due to finish in 60 minutes.  The chair is joehuang. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:53 <openstack> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:55 <longxiongqiu> hello
14:00:55 <openstack> The meeting name has been set to 'tricircle'
14:01:05 <hejiawei> hi
14:01:09 <joehuang> #topic rollcall
14:01:09 <yinxiulin> hi
14:01:13 <joehuang> #info joehuang
14:01:19 <Yipei> #info Yipei
14:01:19 <dongfeg_1> #info dongfeng
14:01:20 <longxiongqiu> #info longxiongqiu
14:01:25 <hejiawei> #info hejiawei
14:01:34 <yinxiulin> #info xiulin
14:01:55 <zhiyuan> #info zhiyuan
14:02:11 <joehuang> #topic feature implementation review
14:02:44 <joehuang> let's have a brief review on our feature implementation, please describe in short
14:03:10 <RongHui11> hello
14:03:22 <joehuang> hi, ronghui
14:04:34 <dongfeg_1> for me, 1) the async job api is still under test, when apply multiple filters it will fail. 2) doc for async job api also needs modification.
14:04:41 <Yipei> for lbaas, already complete the operation of creating lb
14:05:00 <Yipei> try creating a listener for lb
14:05:07 <zhiyuan> (1) flat network support is merged, but a patch need to be added to support the same physical network in different regions (2) multi-gateway support is updated to create local network/subnet before VM creation
14:06:39 <joehuang> to zhiyuan, for same physical network, I just find that we can simply catch the exception FlatNetworkInUse in central Neutron, and then do nothing, just let the local Neutron to judge again
14:07:43 <joehuang> ok, Dongfeng, functest is needed for API
14:08:17 <RongHui11> the network of longqiu is not good and he have updated a very simple qos patch, after merged，i will update others patch one by one.
14:08:30 <joehuang> You can refer to pod functest, it's using restful request to do functest
14:08:40 <dongfeng_1> ok, got it.
14:08:49 <hejiawei> for me, spec for metering needs update. this week I'm not update it due to my math test. It will be update in this weekend.
14:09:23 <joehuang> ok
14:10:16 <joehuang> good
14:10:44 <joehuang> the milestone pike-1.5 will be on May.2
14:11:35 <joehuang> so if your feature plans to land in pike-1.5, please update in time, so that review can be done ASAP
14:11:45 <zhiyuan> to Joe, I am afraid that if the database record can be correctly restored.
14:12:50 <zhiyuan> though we can create both external networks, when query one of the network, the physical network field may be missing
14:12:54 <joehuang> FlatNetworkIn use only happen for physical network name duplication
14:14:10 <joehuang> need to dig into in more detail
14:15:13 <joehuang> For QoS spec
14:15:25 <zhiyuan> I am afraid that we can not associate one physical network with more than one network, we ignore the exception, at the same time the necessary database record is also missing
14:15:26 <joehuang> I found that there is misunderstanding on the comment
14:16:44 <joehuang> there is no existing async job to do QoS disassociation
14:18:22 <joehuang> @XiongQiu?
14:19:48 <joehuang> xiongqiu, are you online?
14:19:57 <longxiongqiu> yes
14:20:32 <joehuang> do you mean dissassociation need no async. job
14:21:58 <joehuang> for network-qos disassociation, in sync way or async way? for port-qos disassociation, it'll be done in sync or async way?
14:22:04 <longxiongqiu> i mean disassociation will be done by update network async job
14:22:14 <joehuang> I don't understand your comment
14:23:35 <joehuang> then how about port
14:23:40 <longxiongqiu> Openstack qos API conference dont provide qos disassociation API，it is supported by network API
14:24:29 <joehuang> in update_network aync job https://github.com/openstack/tricircle/blob/master/tricircle/common/xrpcapi.py#L100
14:25:31 <longxiongqiu> The implementation is in my new patch
14:25:48 <joehuang> ok, will review your patch
14:25:58 <longxiongqiu> i have tested，it is successful
14:26:03 <longxiongqiu> ok
14:26:17 <joehuang> for port, it'll be don in sync way or async way
14:26:42 <joehuang> I meant the port-qos disassociation
14:29:09 <longxiongqiu> it is similar with network in Neutron source code ， so it use async job as network
14:29:33 <joehuang> no port_update async job
14:30:01 <longxiongqiu> i will add in new patch and test
14:30:37 <joehuang> and for port, it's clear where the pod is, so sync disassociation is enough
14:32:12 <joehuang> for async job, it's mainly for long run task, and make it more reliable
14:32:38 <joehuang> if sync operation is ok, it's better to use sync operation
14:32:42 <longxiongqiu> If it use sync way，i just add some code in update port interface， i will consider the two ways
14:33:02 <joehuang> @zhiyuan, your idea?
14:34:03 <joehuang> and on the other hand, async job is to avoid deadlock
14:34:26 <joehuang> between local plugin and central plugin, if deadlock operation is possible
14:35:13 <zhiyuan> I think the above approach is fine, we just handle policy disassociation in network/port update.
14:36:10 <zhiyuan> just add a new part of code in network/port udpate to process policy disassociation
14:36:12 <joehuang> please have further investigation and verification offline
14:36:30 <longxiongqiu> yes
14:37:02 <joehuang> #topic weekly meeting time
14:37:22 <joehuang> hello, OSIC event also brings impact in tricircle
14:37:47 <joehuang> some contributors from OSIC will have to work on other area
14:38:38 <joehuang> so we can reschedule the weekly meeting, it's inconvenient for us to have a meeting at current time
14:39:04 <joehuang> what's your proposal?
14:39:37 <dongfeg_1> do you mean the intel developers will not review our patches?
14:40:25 <joehuang> they will reduce effort for they may have new job responsibility
14:41:37 <dongfeg_1> it's a pity, more reviewers will accelerate the pace. I always expect there will be more comments.
14:41:50 <joehuang> for sure
14:42:44 <dongfeg_1> for me 9:00 PM is ok
14:43:08 <RongHui11> 9：00 PM in Wednesday
14:43:13 <hejiawei> it's ok to me
14:43:17 <joehuang> you mean restore to previous time slot
14:43:23 <dongfeg_1> yes,
14:43:25 <RongHui11> +1
14:43:29 <dongfeg_1> +1
14:43:29 <joehuang> ok
14:43:44 <WuCheng> +1
14:43:52 <RongHui11> 10:00 pm is a little late for us
14:43:56 <yinxiulin> +1
14:43:56 <Yipei> +1
14:43:58 <hejiawei> +1
14:43:59 <longxiongqiu> agree
14:44:49 <zhiyuan> the time is fine for me
14:44:52 <joehuang> OK, I'll create a new doodle, and send in the mail-list, please vote in the doodle
14:45:04 <dongfeg_1> ok
14:45:12 <joehuang> #topic open discussion
14:45:24 <joehuang> any topic?
14:45:52 <dongfeg_1> no for me
14:45:53 <RongHui11> no for me
14:45:57 <zhiyuan> no
14:46:01 <Yipei> no
14:46:01 <yinxiulin> no
14:46:40 <hejiawei> no
14:46:54 <joehuang> ok, thank you for attending the meeting
14:47:01 <joehuang> #endmeeting