01:00:52 #startmeeting tricircle 01:00:53 Meeting started Wed Jun 7 01:00:52 2017 UTC and is due to finish in 60 minutes. The chair is joehuang. Information about MeetBot at http://wiki.debian.org/MeetBot. 01:00:54 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 01:00:56 The meeting name has been set to 'tricircle' 01:01:31 #topic rollcall 01:01:37 #info joehuang 01:01:39 #info dongfeng 01:01:42 #info Yipei 01:02:07 #info xiulin 01:02:47 #topic Pike-2 01:03:24 hello, according to the plan, we'd like to tag the pike-2 on Jun.9 01:04:03 #info zhiyuan 01:04:35 is there any challenge for the pike-2? 01:05:42 actually no for me. my pagination feature can't catch up with it. 01:05:58 to Dongfeng, thanks 01:06:16 I saw patches submitted for the pagination 01:06:40 yes, it need further test. 01:06:53 still working on lbaas, but i think i can finish it soon, still have some problems with health management 01:07:25 will discuss it with you later 01:07:39 Is there any new feature need to implement for LBaaS? 01:08:36 not sure yet, maybe health management needs some code 01:08:51 Good 01:09:24 Tricircle not only can work with multi-region deployment, but also nova cells v2 01:09:51 for nove cells v2, LBaaS will only work with one Nova instance, it'll be much easier 01:10:45 currently you are working on LBaaS in multi-region ( multi-nova instances) scenario 01:11:01 no still one region 01:11:09 no, still one region 01:11:25 still one region, ok 01:12:32 ok, so we'll still tag the pike-2 on Jun.9, if patches can not land before this, will postpone to next cycle 01:12:56 #info tag pike-2 on Jun.9 01:14:18 By the way, there will be a demo in OPNFV Beijing SUmmit next week 01:15:01 Achieving Global Scale VNF High Availability with Multi-Site (Huawei Technologies, Orange, China Mobile) http://events.linuxfoundation.org/events/archive/2017/opnfv-summit/extend-the-experience/opnfv-poc-zone 01:15:36 next week I'll be attending OPNFV Beijing summit. 01:15:55 can someone take the responsibility to chair the weekly meeting? 01:17:26 I have submitted several bug fixes, can see if we can land them in pike-2 01:17:55 ok, will review these patches 01:18:16 ok, let me have a try 01:19:00 thank you zhiyuan 01:21:43 community is working on storyboard experiment to see if launchpad can be migrated to storyboard 01:22:31 #link https://storyboard-dev.openstack.org/ 01:22:41 #link https://storyboard.openstack.org/ 01:22:51 #link https://storyboard-blog.io/ 01:23:00 #link http://docs.openstack.org/infra/storyboard/ 01:25:19 How do you think about using storyboard to replace launchpad? 01:27:06 hmm, need to go through this page "https://storyboard-blog.io/" to know more about storyboard 01:27:33 It's not so intuitive like launch pad 01:27:49 and not sure how well it work with bug tracking 01:28:04 and notification service 01:29:41 I will play with it and be more familiar with it. 01:30:25 ok, feedback are welcome in next weekly meeting 01:30:31 I find that we can login with our ubuntu open ID 01:30:56 the one we use to login review.openstack.org 01:31:02 #info feedback for storyboard and opinion for migration 01:31:29 #topic open discussion 01:31:51 other topic? 01:31:58 yes 01:32:28 please 01:32:52 for lbaas, after creating the lb, the lb and the amphora can be created successfully, but the health management cannot ping the amphora in the management network 01:32:56 i checked 01:33:32 and found that the vlan tag of the interface o-hm0 of br-int is changed to 4096 01:34:02 instead of the one which is the same to the dhcp port of the health management network 01:35:20 before creating the lb, the o-hm0 interface and the tap of the dhcp port is added to br-int 01:35:32 and share the same vlan tag 01:36:02 so the dhcp port is ping-able in default network namespace 01:36:08 yipei, you mean ovs-vswitch will modify the vlan tag o-hm0 from a valid value to 4096? 01:36:14 so the o-hm is not in the management network, it should be, isn't it? 01:36:16 yes 01:36:30 to zhiyuan, yes 01:37:37 to joehuang, o-hm is a port added to the br-int, its ip and mac address are a port in management network 01:38:33 joehuang: if you or anyone have any questions about storyboard in the #storyboard channel. 01:38:50 ok 01:39:04 to diablo_rojo_phon 01:41:10 strange, if ovs-vswitch will set the vlan tag of a unrecognized interface to 4096, the original LB without Tricircle will have the same problem 01:41:48 also, i checked the plugin.sh in octavia, there is not code for setting the vlan tag 01:42:56 so i think maybe there is something wrong when creating the o-hm0 interface manually. but i do i following the plugin.sh 01:43:11 #link https://github.com/openstack/octavia/blob/master/devstack/plugin.sh#L289 01:43:37 how about the plugin for non-tricircle environment? 01:44:28 not tested, but should be OK, i will have a try 01:44:53 yes, and check the ovs flow 01:45:11 after running the line289, the vlan tag of o-hm0 port is 4096 01:45:33 this is incorrect if VLAN tag is 4096 01:45:37 which caused the problem discussed last week 01:46:16 did you apply all the parameter? 01:46:45 I see this "-- set Interface o-hm0 external-ids:iface-id=$MGMT_PORT_ID" 01:46:50 yes, all the parameters, without any modification 01:47:06 so I guess the vlan tag is set by ovs-agent 01:50:11 How about the lb management network segment id in local neutron 01:53:22 you mean whether the network id is in local neutorn? 01:54:24 vlan id of the network in local neutron 01:56:22 to joehuang, i checked, the management network in the local neutron is vxlan 01:56:26 or vxlan id 01:56:44 yes, what's the vxlan id if it's vxlan network 01:56:47 segment_id is 1086 01:58:28 for vxlan network, in the same node, there is one local vlan used inside that compute node 01:58:35 is the port for o-hm0 created in local Neutron? 02:00:01 need to get port in local neutron too 02:00:12 except get network 02:00:14 to zhiyuan, no, it is created in central neutron 02:00:57 ok, time is up, please move to #openstack-tricircle for further discussion 02:01:01 ok 02:01:03 #endmeeting