14:04:46 <fdegir> #startmeeting 3rd Party CI 14:04:46 <collabot`> Meeting started Wed Nov 30 14:04:46 2016 UTC. The chair is fdegir. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:04:46 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic. 14:04:46 <collabot`> The meeting name has been set to '3rd_party_ci' 14:05:08 <fdegir> #topic bifrost status 14:05:23 <fdegir> yolanda: hwoarang: I think you have things to say 14:05:24 <yolanda> so from my side, i started to test openstack-ansible with bifrost 14:05:37 <fdegir> #info yolanda started to test openstack-ansible with bifrost 14:05:45 <yolanda> i could make it work but with some manual workarounds, as we were hitting a problem with bifrost 14:06:03 <yolanda> openstack-ansible documentation http://docs.openstack.org/developer/openstack-ansible/install-guide/app-config-test.html#test-environment-config 14:06:08 <fdegir> #info it works but with some manual workarounds, as we were hitting a problem with bifrost 14:06:08 <Julien-zte> #info Julien 14:06:15 <yolanda> shows how to deploy using multiple vlans/nics 14:06:17 <fdegir> #link http://docs.openstack.org/developer/openstack-ansible/install-guide/app-config-test.html#test-environment-config 14:06:23 <yolanda> and bifrost currently only supports one 14:06:30 <fdegir> #info the above link shows how to deploy using multiple vlans/nics 14:06:39 <fdegir> #info and bifrost currently only supports one 14:06:43 <yolanda> so what i did is just the deploy with bifrost, and then creating vlans and bridges with ip commands 14:06:59 <yolanda> but that does not persist, is cleaned on each network restart, as glean is wiping those 14:07:16 <fdegir> #info the test involved deploying with bifrost and then creating vlans and bridges with ip commands 14:07:16 <yolanda> i talked with Julia, and she is working on a change in bifrost, to allow it to manage multiple nics/vlans 14:07:19 <fdegir> #info but that does not persist, is cleaned on each network restart, as glean is wiping those 14:07:32 <Julien-zte> "bifrost currently only supports one" what does one eman? vlan? 14:07:38 <yolanda> nic or vlan 14:07:43 <ChrisPriceAB> vlan per nic I believe 14:07:44 <fdegir> #info Julia is working on a change in bifrost, to allow it to manage multiple nics/vlans 14:07:50 <yolanda> either one nic or one vlan 14:08:00 <ChrisPriceAB> I was wrong. :) 14:08:11 <fdegir> as usual :) 14:08:13 <yolanda> it is configured to have just one nic/vlan for pxe boot, one nic/vlan for network 14:08:16 <yolanda> heh :) 14:08:20 * ChrisPriceAB sulks off 14:08:36 <fdegir> #info it is configured to have just one nic/vlan for pxe boot, one nic/vlan for network 14:08:39 <yolanda> so Julia shall be working on that, but is a complex change 14:09:10 <fdegir> so yolanda, have you been able to deploy openstack with openstack-ansible after you manual stuff? 14:09:11 <ChrisPriceAB> would be nice to be able to inject a config file 14:09:28 <fdegir> or you always experience the cleanup? 14:09:28 <yolanda> fdegir, got it working somehow yes. it was failing on storage 14:09:56 <fdegir> and this is on baremetal? 14:10:07 <yolanda> yes, i tried on baremetal 14:10:23 <ChrisPriceAB> 1 control 1 compute? 14:10:24 <yolanda> it connects by ssh, and deploys all services using lxc 14:10:47 <yolanda> 1 control and 1 compute yes, seems i needed a node for storage, or tweak parameters somehow 14:10:54 <yolanda> that's problem i was facing lately 14:11:25 <fdegir> yolanda: do you think you can write up a simple readme in releng/prototypes/openstack-ansible for the manual steps etc? 14:11:47 <fdegir> so we don't have to bug you inc ase if others want to give it a try? 14:12:04 <yolanda> yep, i can write what i had, but is not yet 100% working 14:12:12 <yolanda> but can work as a starting point 14:12:13 <fdegir> no hurries 14:12:25 <fdegir> whenever you have something usable by others, that would be fine 14:12:33 <fdegir> what do you think about virtual deployments? 14:12:44 <fdegir> since we don't have enough baremetal machines 14:12:52 <fdegir> so others can try with vms 14:13:02 <yolanda> i have not tried but shall be something similar as we did with puppet-infracloud 14:13:07 <yolanda> create the vms, deploy with ironic, etc 14:13:19 <fdegir> ok 14:13:21 <yolanda> and again, there will be a need to manually create the missing vlans and bridges 14:13:35 <fdegir> which needs readme :) 14:13:51 <hw_wutianwei> I deploy openstack with openstack-ansible on vms node launching manually, it is not successful everytime 14:14:26 <yolanda> hw_wutianwei, you could share your expertise there as well. I was hitting config problems with the storage 14:14:35 <yolanda> so if you can share your deployment file, will be useful 14:15:16 <hw_wutianwei> I use three nodes1 control and 1 compute and 1storage 14:16:36 <fdegir> I suppose we can move on 14:16:37 <yolanda> ah that was my problem, trying to deploy without the additional storage 14:16:55 <Julien-zte> hw_wutianwei, why a separate storage node? same reason, blocked by NIC? 14:16:58 <yolanda> i could check in openstack-ansible if that's possible as well 14:17:55 <hw_wutianwei> I accord to the test example 14:18:14 <hw_wutianwei> http://docs.openstack.org/developer/openstack-ansible/install-guide/app-config-test.html#test-environment-config 14:20:00 <hw_wutianwei> I accord with the test example file 14:20:07 <fdegir> next one in the agenda is bifrost on suse 14:20:09 <fdegir> hwoarang fixed bifrost on suse 14:20:11 <fdegir> finally... :) 14:20:20 <fdegir> hwoarang: anything you want to add? 14:20:34 <hwoarang> so i need to merge two more patches to diskimage-builder 14:20:49 <hwoarang> because i have to fix the slave manually whenever bifrost updates diskimage-builder 14:21:05 <hwoarang> and then i think i am done in regards to suse+bifrost. at least for the major bits 14:21:30 <fdegir> #info 2 more patches are needed to be merged for diskimage-builder 14:22:05 <fdegir> #info once that is done, work with bifrost on suse is deemed to be complete 14:22:17 <fdegir> hwoarang: and then to openstack-ansible on suse? 14:23:05 <fdegir> I take this as yes 14:23:08 <hwoarang> yes 14:23:14 <hwoarang> but haven't started yes 14:23:18 <hwoarang> *yet 14:23:30 <fdegir> Julien-zte: your turn 14:23:34 <fdegir> what are you up to lately? 14:24:01 <Julien-zte> no progress recently. most of the resources work for an inner project temporary 14:24:05 <Julien-zte> sorry for this 14:24:14 <fdegir> np 14:24:25 <Julien-zte> it nearly finished 14:24:35 <fdegir> jmorgan1: anything from your side? 14:24:45 <fdegir> jmorgan1: when are you going to dump us from intel lab? 14:25:20 <hwoarang> :( 14:25:33 <fdegir> I take this as no, he is not going to do that :) 14:25:49 <Julien-zte> any new resources required? 14:26:06 <fdegir> Julien-zte: we will need when jmorgan1 starts his lab migration 14:26:43 <fdegir> Julien-zte: or even before so we don't stop our bifrost 3rd party ci jobs on jenkins 14:26:51 <fdegir> Julien-zte: do you have some? 14:26:55 <Julien-zte> I can provide by 2-3 weeks. but I'm not sure the internet link speed from your contry is acceptable 14:27:27 <fdegir> Julien-zte: we can give a try with 1 machine and see how it goes 14:27:44 <Julien-zte> The lab construction will be finished by 1-2 weeks, then I have one week to prepare 14:28:07 <fdegir> Julien-zte: ok, will ping you then 14:28:16 <Julien-zte> good 14:28:24 <fdegir> from my side 14:28:36 <fdegir> I've been working on setting up 3rd party ci for ODL 14:28:42 <Julien-zte> If the infrasture is ready, I can allocate some account for test 14:29:16 <fdegir> #info we had odl gerrit and opnfv jenkins connected so we can trigger jobs on our jenkins for the odl patches 14:29:40 <fdegir> #info work in progress to settle the deployment with apex and functest 14:30:02 <fdegir> that's all from me 14:30:08 <fdegir> anyone else wants to add anything? 14:30:25 <Julien-zte> how many server do you need fdegir? 14:30:40 <fdegir> 3 for CI, 3 for development 14:30:45 <fdegir> a full pharos pod 14:31:06 <Julien-zte> baremetal, no openstack deployed? 14:31:10 <fdegir> but once the intel lab migration is done, we can return some back 14:31:13 <fdegir> baremetal 14:31:18 <Julien-zte> good 14:31:35 <fdegir> we just need OS - ubuntu, centos, and suse 14:31:45 <Julien-zte> get it 14:32:02 <fdegir> ChrisPriceAB: you want to add anything 14:32:09 <fdegir> like "great job guys, keep up" :) 14:32:17 <Julien-zte> haha 14:32:46 <fdegir> then we are done for they day 14:32:54 <Julien-zte> it's really great 14:32:56 <fdegir> thanks to you all and talk to you in 2weeks 14:32:58 <Julien-zte> thanks all 14:33:00 <fdegir> #endmeeting