01:01:08 #startmeeting tricircle 01:01:09 Meeting started Wed May 31 01:01:08 2017 UTC and is due to finish in 60 minutes. The chair is joehuang. Information about MeetBot at http://wiki.debian.org/MeetBot. 01:01:10 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 01:01:12 The meeting name has been set to 'tricircle' 01:01:21 hi 01:01:38 Many of us just had public holidays on Monday and Tuesday 01:01:58 So maybe not so many people will be shown in the meeting, it's ok 01:02:08 let's have open discussion 01:03:17 ok, for lbaas, when creating health management interface, i fail to allocate the ip to it by using dhclient 01:03:23 can i configure it manually? 01:04:07 ok, I still try to implement the pagination for routing list, because I'm not familiar with sqlalchemy, it will need some time 01:04:26 have you checked the dhcp name space 01:04:39 whether the dhcp namespace for the network is working 01:05:09 to Dongfeng, thanks a lot 01:05:18 :) 01:06:28 if you want to allocate ip to the instance via dhcp, the dhcp name space should be there 01:06:35 the dhcp network namespace exists 01:07:12 i can see a tap for the dhcp port 01:07:20 ok 01:07:21 in the network namespace 01:09:11 I remember that in octavia devstack plugin, there is some scripts for the health management interface 01:10:09 #link https://github.com/openstack/octavia/blob/master/devstack/plugin.sh#L274-L312 01:10:11 yes, i create the interface by following the plugin.sh 01:10:44 sudo dhclient -v o-hm0 -cf $OCTAVIA_DHCLIENT_CONF 01:11:19 yes, that is the command 01:11:37 /etc/octavia/dhcp/dhclient.conf 01:11:53 the network is dhcp enabled or disabled? 01:12:24 the dhcp agent has been bound or not? 01:13:03 I am afraid that in Tricircle there is some processing in DHCP will make o-hm not work normally 01:13:27 i think it is enabled, because i create the management port without specifying a fixed ip, and i also successfully boot a vm 01:14:14 i think it the source of the amphora reboot, the health management cannot work well 01:14:45 how to check whether the dhcp agent has been bound or not 01:15:05 have you installed tricircle local plugin in local Neutron? 01:16:02 yes, i installed tricircle and octavia together with devstack, but comment the lines in plugin.sh of octavia, which creates mgmt networks 01:16:05 openstack network agent list/show/etc 01:16:34 so that i can create the management network manually 01:18:14 the management network is created from central Neutron or from local neutron? 01:19:24 sorry i am late. to yipei, you can check whether there is a dhcp namespace for the management network 01:19:39 dhcp namespace is there 01:20:32 then can you ping the dhcp IP inside the VM? 01:20:49 to joehuang, i execute openstack network agent list, all the states of agents are UP 01:21:15 to zhiyuan, i can ping the dhcp IP in the dhcp network namespace 01:21:33 yes, you can also boot one VM in that mngt network to see if the VM can get IP 01:22:06 If you ping inside dhcp name space, you just ping loopback 01:23:33 yes, i tried, the vm can be successfully booted. i booted a vm, because if i do not boot one, there is not dhcp network namespace 01:23:58 i can ping loopback, too 01:24:10 can the dhcp IP be ping-able inside vm 01:25:37 to joehuang, but the o-hm0 interface is in default network namespace 01:27:26 to joehuang, i tried 01:27:38 the dhcp ip is ping-able inside the vm 01:29:28 to yipei, in the host running the dhcp agent, run "ps axu | grep dnsmasq" 01:29:31 0 01:30:29 check if the IP of o-hm0 is inside this file: /opt/stack/data/neutron/dhcp/$SOME_ID/addn_hosts 01:30:46 if vm can get IP, I think dhcp is normal 01:30:50 i am not sure $SOME_ID is network id or subnet id 01:31:10 you can see from the result of "ps" 01:31:39 nobody 1199 0.0 0.0 54608 2664 ? S 5月30 0:03 /usr/sbin/dnsmasq --no-resolv --keep-in-foreground --no-hosts --bind-interfaces --pid-file=/var/run/NetworkManager/dnsmasq.pid --listen-address=127.0.1.1 --cache-size=0 --conf-file=/dev/null --proxy-dnssec --enable-dbus=org.freedesktop.NetworkManager.dnsmasq --conf-dir=/etc/NetworkManager/dnsmasq.d 01:31:39 nobody 29318 0.0 0.0 49988 2888 ? S 08:58 0:00 dnsmasq --no-hosts --strict-order --except-interface=lo --pid-file=/opt/stack/data/neutron/dhcp/e89bb941-2493-4146-8986-a2ea5990218c/pid --dhcp-hostsfile=/opt/stack/data/neutron/dhcp/e89bb941-2493-4146-8986-a2ea5990218c/host --addn-hosts=/opt/stack/data/neutron/dhcp/e89bb941-2493-4146-8986-a2ea5990218c/addn_hosts --dhcp-optsfile=/opt/stack/data/neutron/dhcp/e89bb941-2493-4146-8986- 01:31:40 a2ea5990218c/opts --dhcp-leasefile=/opt/stack/data/neutron/dhcp/e89bb941-2493-4146-8986-a2ea5990218c/leases --dhcp-match=set:ipxe,175 --bind-interfaces --interface=tap99e6b6d7-4a --dhcp-range=set:tag0,192.168.0.0,static,86400s --dhcp-option-force=option:mtu,1450 --dhcp-lease-max=256 --conf-file= --domain=openstacklocal 01:31:43 stack 31234 0.0 0.0 15960 1020 pts/36 S+ 09:30 0:00 grep --color=auto dnsmasq 01:32:10 yipei, do you have this port locally in local Neutron, or the port is only in central Neutron? 01:32:37 i only create the port in central neutorn 01:32:48 so no port in local neutron 01:32:53 you need to get the port in local neutron first 01:33:10 only if there is port in local neutron, then dhcp will work 01:34:07 but octavia uses central neutron to manage amphora as well as the management network 01:34:09 after creating a port in central Neutron, you can use "openstack port show $central_port_id" to create the port in local Neutron 01:34:58 "openstack --os-region-name $local-region port show $central_port_id" 01:35:14 Yipei, do you know how to address this issue now? 01:35:22 yes 01:35:25 i see 01:36:26 ok 01:36:35 for some information to share 01:37:28 two applications running in OPNFV summit multisite now all work 01:38:29 both vIMS(voice) and JITSI(video conference) can run in multi-site environment with site level redundancy 01:39:07 through the networking capabilities, these two applications can achieve site level high availability 01:39:47 vIMS is via Triricle cross site L3 networking 01:40:00 JITISI is via cross site L2 networking 01:40:45 #info through the networking capabilities of tricricle, vIMS and JITSI can achieve site level high availability in OPNFV multisite environment 01:41:49 another news is that Neutron + Tricircle can work well with Nova cellsv2 01:42:07 #link http://lists.openstack.org/pipermail/openstack-dev/2017-May/117599.html 01:42:38 other topic to discuss? 01:43:20 to zhiyuan please review the QoS spec :) 01:43:43 Ronghui, please communicate with ZHiyuan about the comment 01:43:53 ok 01:44:19 and some patches submitted should include test cases too 01:44:32 ok will update in next version 01:44:46 I have talked to Xiongqiu about the "create_qos_policy" job 01:45:23 yeah, please update in time to make it being integrated fast 01:46:06 if no other topic, let's conclude the meeting 01:46:33 At first he was going to create the local qos policy in both "network_update" and "create_qos_policy" jobs, so I suggest to only create qos policy in "create_qos_policy" job, then also update network to bind qos policy in this job 01:48:14 please refer to https://docs.google.com/presentation/d/16Qfm3D1yhrW31ZEVRYea8mcyKI06fQ3a6Okk4VhFGkQ/ for OPNFV multisite environment 01:49:16 #info refer to https://docs.google.com/presentation/d/16Qfm3D1yhrW31ZEVRYea8mcyKI06fQ3a6Okk4VhFGkQ/ for OPNFV multisite environment 01:49:26 other topic? 01:49:33 to joehuang and zhiyuan, i tried, still does not work, can we discuss it in tricircle channel after the meeting? 01:49:43 ok 01:49:44 ok 01:49:51 no for me 01:50:02 no for me 01:50:05 no for me 01:50:25 #endmeeting