01:04:41 #startmeeting tricircle 01:04:42 Meeting started Wed Sep 27 01:04:41 2017 UTC and is due to finish in 60 minutes. The chair is zhiyuan. Information about MeetBot at http://wiki.debian.org/MeetBot. 01:04:43 Useful Commands: #action #agreed #help #info #idea #link #topic #startvote. 01:04:45 The meeting name has been set to 'tricircle' 01:05:44 at first, just to remind that we have merged an urgent patch for the smoke test: https://review.openstack.org/#/c/507340/ 01:05:44 patch 507340 - tricircle - [Urgent] Add domain infor in smoke test (MERGED) 01:06:31 Last week, Song, zeng you, Yaguang, Xiaohai have join TriCircle. They have finished the compile of the source code in their local envionment. 01:06:32 if your patch failed to pass the smoke test, please do a rebase or recheck 01:07:15 got it and thank zhiyuan so much for fix this problem 01:08:08 to ztang, so they have started the devstack environment with tricircle? 01:08:17 yes 01:08:47 our szzt team are working on the qos test 01:09:12 cool, any problem found during the environment setup? 01:09:28 help to complete the qos smoke test :) 01:09:30 because of the fire wall it not ease to environment setup 01:09:39 Song, please introduce the whole works of our team in TriCircle in the past week. I hope everyone can have a specific module division in this project. 01:09:41 hi,I am xiaohan zhang from szzt,nice to meet everyone! 01:10:41 we have set up the dev environment and plan to finish all the qos test case this week. 01:12:01 Very well 01:12:22 to song, there are two parts of works for the smoke test. (1) modify devstack/plugin.sh to automatically setup the environment with qos enabled (2) write a yaml file to define the test 01:13:03 yes,thank you for help. we have read the source code of the tricircle test. 01:14:10 but we have a little question about the test environment. 01:14:20 please 01:14:39 it is the test environment have fix machine. 01:14:59 to zhiyuan, four new developers have joined to finish the QoS module. 01:16:33 Song is in the Top 50 contributors for CEPH in the world. 01:16:48 sorry for late, (can't connet to freenode this morning) 01:16:59 great, this can significantly speed up the development! 01:17:23 we do not clear about the test environment such as the ip and net,are these all changeless. 01:17:44 to ztang yes i am working on ceph in before new to openstack. 01:18:05 openstack CI will schedule some virtual machines to run the test 01:18:08 but out team have other peoples all good at in openstack. 01:18:41 to zhiyuan oh got it.thank you ! 01:20:20 we found the doc for tricircle devlop is a little few, do we plan to add more doc for develop 01:20:26 you can refer to tricircle/tempestplugin/gate_hook.sh to see how to obtain the IP, if necessary 01:21:32 to zhiyuan i will set if after the meeting. thanks again. 01:22:19 yes, we lack a doc for how to write yaml definition, but we have a spec: https://github.com/openstack/tricircle/blob/master/specs/pike/smoke-test-engine.rst 01:22:37 which can be linked somewhere in our doc 01:22:53 i have read the specs.to write yaml it is really a good idea. 01:25:41 one thing to mention, currently the smoke test only validate whether the resources have been correctly created in local neutron, we don't login the VM to ping other VMs, to test real data plane connectiviry 01:25:52 s/connectiviry/connectivity 01:26:18 got it 01:27:52 so you can see that the "validate" task just list resources in local neutron then check their attributes 01:29:05 do we have the test interface to do that? if not add one for all the test case to use. 01:30:35 no, currently we don't have such interface. need to see how this is implemented in tempest 01:32:04 oh,got it. i think the interface it is need to add. maybe in the local neuron is ok. but it is not really ok. 01:33:37 yes, that's possible 01:37:09 i also have a little question about the pr.it is that ok to put the test code and source code in diff pr? 01:38:58 ?? tricircle repos contains both source and test 01:39:15 oh got it because we have diff people to finish the work. 01:40:20 you can put the feature implementation in one patch, and test in another patch that depends on the former one 01:41:17 how about the lbaas feature, yipei? I see your reply in the mail list, so michael's reply doesn't solve your problem? 01:41:39 to zhiyuan oh thank sir.got it. 01:41:45 no, it doesn't 01:42:44 i tried his solution to verify whether the route table is correct 01:43:01 it is correct in my env 01:45:43 how the amphorae image is generated? 01:46:24 it is generated by octavia itself during installation 01:49:18 hmm, seems like the problem of haproxy 01:49:33 very strange 01:49:35 Yipei maybe I mis-understood the scenario. Can you reply with a set of reproduction instructions? 01:50:01 I will try to reproduce it again 01:50:12 ok, sure 01:51:40 to johnsom, in the mailing-list or here? 01:52:15 Mailing list would be best, or etherpad and post the link in the lbaas channel 01:52:34 ok, got it, thank you, johnsom 01:52:49 Sure, I will try to help 01:54:09 i will send a mail in the mailing list 01:55:02 any other topics to discuss? 01:55:11 no 01:55:43 for default security group update, i have write the first version code(not test). But I'm 01:55:43 puzzled about how to test xjob, i'm looking a solution. 01:57:09 if the job fails, you can stop xjob, do some midification, then restart xjob 01:57:20 xjob will re-run the failed job 01:57:22 for default sg update, could you give me some advice(or some example) about how to use xjob and how to test 01:58:42 configure_security_group_rules in xmanager.py is used to sync security group rules 01:59:11 enen 01:59:58 after sg update, you can call the rpc api in common/xrpcapi.py 02:00:21 there is a function with the same name 02:01:22 yeah, I am using this solution 02:01:57 then xjob can be informed to run the job 02:02:43 oh, time's up 02:02:50 yeah, thanks 02:03:20 let's end our meeting, thanks for attending 02:03:45 #endmeeting