14:01:59 <fdegir> #startmeeting Cross Community CI (XCI) Meeting
14:01:59 <collabot> Meeting started Wed Jan 25 14:01:59 2017 UTC.  The chair is fdegir. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:01:59 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:01:59 <collabot> The meeting name has been set to 'cross_community_ci__xci__meeting'
14:02:01 <jmorgan1> hwoarang: no, i don't own or manage that lab, I'm just a user/customer
14:02:08 <fdegir> #topic Rollcall
14:02:16 <fdegir> who is around?
14:02:18 <jmorgan1> #info Jack Morgan
14:02:25 * fdegir sees all of you
14:02:33 <fdegir> #info Fatih Degirmenci
14:02:35 <yolanda> #info Yolanda Robla
14:02:39 <hw_wutianwei> #info Tianwei  Wu
14:02:49 <hwoarang> #info Markos Chandras
14:03:01 <hwoarang> jmorgan1: ok sorry didnt know
14:03:15 <jmorgan1> hwoarang: no problem
14:03:47 <fdegir> #info Fatih Degirmenci
14:03:59 <s_berg> #info Stefan Berg
14:04:26 <fdegir> #topic OpenStack XCI Status (bifrost/openstack-ansible)
14:04:26 <jmorgan1> s_berg: I'll take care of yoru von request today - sorry for delay
14:04:37 <fdegir> any update from anyone?
14:04:37 <NikoHermannsEric> #info Nikolas Hermanns
14:05:03 <s_berg> jmorgan1: Thanks, but no worries - been down and out due to a bad case of the flu the last two weeks. Just crawling back to work this week. :)
14:05:04 <hw_wutianwei> i have create a job for bifrost and osa,  planning to run in huawei-virtual6.
14:05:21 <hw_wutianwei> fdegir: could you change the tag of this virtual into infra-osa-virtual-xenial?
14:05:46 <fdegir> hw_wutianwei: which job you are talking about? - I'm a bit behind with the reviews
14:05:52 <fdegir> late^
14:06:33 <fdegir> hw_wutianwei: can you put the link to job or gerrit patch please?
14:06:34 <hw_wutianwei> I add the job to run bifrost and then run osa-deploy
14:06:36 <hw_wutianwei> https://gerrit.opnfv.org/gerrit/#/c/27375/
14:07:09 <fdegir> hw_wutianwei: it fails on jenkins verificaiton and I see some comments there
14:07:33 <fdegir> hw_wutianwei: can we let people try to run stuff manually based on what you intend to run by job?
14:08:03 <fdegir> hw_wutianwei: and we fix the rest once you have reviews addressed there?
14:08:15 <yolanda> hw_wutianwei, the latest issue i ran, is the conflict between ntp/chrony, and that for some reason, the lxc_attach command was not found
14:08:17 <hw_wutianwei> this job verify the bifrost and osa
14:09:08 <fdegir> hw_wutianwei: before we attempt to verify upstream patches, shouldn't we make sure it works?
14:09:14 <hw_wutianwei> yolanda: in my opinion, the osa will install the lxc when deploying
14:09:19 <fdegir> hw_wutianwei: or else we will send some wrong feedback to upstream
14:09:26 <hw_wutianwei> I will fix this
14:09:46 <fdegir> hw_wutianwei: I should perhaps put my comment there but let me put it here as well
14:10:02 <fdegir> hw_wutianwei: until we make it work locally, we shouldn't even create the job against openstack gerrit
14:10:23 <yolanda> hw_wutianwei, i have not seen any place where OSA installs lxc-common package...
14:11:14 <fdegir> #info hw_wutianwei uploaded a patch to create bifrost/osa verification jobs
14:11:16 <yolanda> and as we run with the ubuntu-minimal, you could expect that lxc-common is not installed by default, i'd try adding to the list of installed packages on the image
14:11:21 <fdegir> #info The patch is up for review
14:11:27 <fdegir> #link https://gerrit.opnfv.org/gerrit/#/c/27375/
14:11:55 <hw_wutianwei> yolanda: i test in ubuntu 16.04, there are some task install the lxc, i will send the log, and you can chcek
14:12:02 <fdegir> #info There were some issues with osa due to the conflict between ntp/chrony
14:12:21 <yolanda> ok hw_wutianwei , let me see, i tested a pair of times and both failing with the same :(
14:12:25 <fdegir> #info And "lxc_attach command not found" error
14:12:52 <yolanda> i also found some problems with memory exhausted on my server
14:12:55 <hw_wutianwei> i will fix conflict between ntp/chrony
14:13:02 <yolanda> as we spin up 6 vms * 16gb
14:13:13 <hw_wutianwei> maybe  too much vms
14:13:27 <yolanda> i have the proposal that maybe we can have different memory requirements per vm
14:13:29 <fdegir> #info hw_wutianwei will take a look into ntp/chrony conflict
14:13:34 <yolanda> so for example, we don't need 16gb on the jumphost
14:13:56 <yolanda> we could amend the bifrost script to support that, same thing as we did for specifying independent names on the vms, what do you think?
14:14:28 <fdegir> #info Memory requirement per vm might be an issue for people - yolanda is attempting to lower memory allocation
14:14:34 <hw_wutianwei> yolanda: that is great.
14:14:46 <fdegir> hw_wutianwei: what about your earlier patch?
14:14:51 <fdegir> hw_wutianwei: https://gerrit.opnfv.org/gerrit/#/c/26865/
14:14:52 <yolanda> i can send some proposal to bifrost for it
14:15:18 <fdegir> hw_wutianwei: do you need help?
14:15:43 <jmorgan1> I think the target is to get this working on a single server for vPOD that doesn't require a large amount of hardware resources
14:15:44 <fdegir> hw_wutianwei: others might help if you like by sending new patches to same change?
14:15:45 <hw_wutianwei> fdegir: this patch still have some bugs need fix
14:16:15 <hwoarang> this patchset is huge. it would be helpful to split it into smaller independent changes
14:16:35 <hwoarang> make them all live under the same 'topic'
14:16:48 <fdegir> I think we can first make it work for us manually
14:16:53 <fdegir> like how we did it with bifrost
14:17:00 <fdegir> and the jobs etc can wait a little longer
14:17:25 <fdegir> if we can get multiple people to run stuff manually, that would be a good sign to put effort to create jobs etc
14:17:36 <fdegir> and I agree to what jmorgan1 and hwoarang say
14:18:00 <fdegir> we need to make it work on a single server vpod so we can work on stuff without needing to have additional resource
14:18:01 <yolanda> i tried to run lowering the vms to 8gb because with 16gb it was very slow to me. But i could not end the testing due to that lxc problems
14:18:27 <yolanda> also some problems in the playbooks, you need to run twice the prepare infra playbook, or ansible cannot connect by ssh
14:18:28 <fdegir> and have smaller patches to develop it incrementally
14:18:49 <hw_wutianwei> i have tested that, it is ok for. so others can test that. if having any question we can solve that
14:19:11 <fdegir> thanks a lot for the patches hw_wutianwei
14:19:17 <fdegir> it's much appreciated
14:19:27 <hw_wutianwei> yolanda: i will fix that problem in next time
14:19:31 <yolanda> yep, great work, hw_wutianwei !
14:19:42 <fdegir> I think we have 3 osa related items to tackle
14:19:53 <fdegir> #info split the first patch to smaller patches
14:20:09 <fdegir> #info make it work on a single machine by finding lower hw requirements (memory etc)
14:20:14 <hw_wutianwei> next week is chinese new year
14:20:14 <fdegir> #info fix ntp/chrony issue
14:20:22 <hw_wutianwei> so i will out of work
14:20:39 <fdegir> hw_wutianwei: are you ok if we send patches to your change?
14:20:50 <fdegir> hw_wutianwei: while you are away?
14:21:17 <hw_wutianwei> yep, i will check that
14:21:28 <fdegir> thanks
14:21:32 <fdegir> are we done with osa?
14:22:03 <fdegir> ok
14:22:08 <fdegir> hwoarang: now you
14:22:17 <hwoarang> right
14:22:18 <fdegir> hwoarang: i see you are keeping jenkins busy
14:22:27 <fdegir> hwoarang: how is it going with bifrost log and other stuff?
14:23:00 <hwoarang> logs should be OK now. I have set a new patch to add one new job to clean up logs for merged patchsets
14:23:11 <hwoarang> and there is also this centos7 issue that i need to fix properly
14:23:18 <hwoarang> asked neutron and ironic people and we will see
14:23:42 <fdegir> #info bifrost log upload/cleanup is coming to an end
14:24:11 <fdegir> hwoarang: do you think if we should wait with voting rights until we learn what's wrong with centos?
14:24:12 <hwoarang> oh and yeah, please review this https://gerrit.opnfv.org/gerrit/#/c/27439/
14:24:15 <fdegir> (apart from it being centos)
14:24:33 <fdegir> #info Issues with bifrost on centos still exist
14:24:46 <fdegir> #info hwoarang contacted neutron/ironic people
14:24:48 <hwoarang> fdegir: i can make it work if I hack it. let me talk to the neutron people and see where the problem really is
14:24:50 <hwoarang> lets wai t a bit
14:24:53 <hwoarang> *wait
14:24:56 <fdegir> agree
14:25:12 <fdegir> #info Voting rights will be delayed until we have more info about centos issues
14:25:39 <fdegir> anyone else wants to talk about osa and/or bifrost?
14:25:49 <fdegir> any progress/attempts/issues/needs?
14:26:16 <fdegir> I take this as no
14:26:22 <fdegir> moving to next topic, which is
14:26:34 <fdegir> #topic OpenDaylight XCI Status
14:26:43 <fdegir> we have NikoHermannsEric with us
14:26:50 <fdegir> who is driving ODL XCI efforts
14:26:55 <NikoHermannsEric> Hey,
14:26:57 <fdegir> thanks for joining NikoHermannsEric
14:27:02 <fdegir> any update from your side?
14:27:05 <NikoHermannsEric> yes
14:27:11 <fdegir> please go on
14:27:29 <NikoHermannsEric> Tim and me are still working on 2 patches to bring automatic snapshot alive
14:27:36 <NikoHermannsEric> I think we will be ready with that soon.
14:27:44 <fdegir> #info NikoHermannsEric and trozet working on 2 patches to bring automatic snapshot alive
14:27:52 <fdegir> #info The work is expected to be completed soob
14:27:56 <fdegir> #undo
14:27:56 <collabot> Removing item from minutes: <MeetBot.ircmeeting.items.Info object at 0x1f8c390>
14:27:58 <fdegir> #info The work is expected to be completed soon
14:28:26 <fdegir> we should perhaps mention the relation between apex snapshots and odl patchset verification so people know what that is
14:28:30 <NikoHermannsEric> In addition some fixes where done in functest so yesterday we had our first successfull run and functest did show green for the current odl master
14:28:54 <fdegir> #info Fixes were done in functest so yesterday we had our first successfull run and functest did show green for the current odl master
14:29:13 <jmorgan1> this is with Apex
14:29:20 <fdegir> jmorgan1: yes
14:29:28 <NikoHermannsEric> in short: an virtual apex deployment from a snapshot of a former install apex will be used as the test_environment
14:30:05 <fdegir> apex will promote working snapshots and store them on artifacts.opnfv.org
14:30:14 <NikoHermannsEric> Apex has an issue right now and cannot be deployed but we need to update the apex snapshot used in the pipeline
14:30:20 <NikoHermannsEric> that work is ongoing
14:30:26 <fdegir> odl patchset verification job will fetch them and bring up the openstack deployment in minutes
14:30:32 <fdegir> so we can provide fast feedback to odl
14:30:49 <fdegir> and catch issues that break opnfv stuff earlier
14:30:52 <NikoHermannsEric> fdegir: thanks for giving more insides
14:30:59 <fdegir> np :)
14:31:15 <fdegir> for the ones who didn't see the job
14:31:19 <fdegir> the link is
14:31:21 <NikoHermannsEric> another addtion it that we don't need to install every test_environment
14:31:29 <fdegir> #link https://build.opnfv.org/ci/view/Cross%20Community%20CI/job/odl-netvirt-verify-virtual-master/
14:31:55 <NikoHermannsEric> so we save a lot of disk ios etc + cpu load. so that we can run 6 virtual deployment on one node without performance issues
14:32:13 <NikoHermannsEric> so the amount of testing resources is QUITE large.
14:32:17 <jmorgan1> do we have any other the ODL CI effort documented anywhere? I mean how to set up
14:32:19 <fdegir> #info another addtion it that we don't need to install every test_environment
14:32:27 <fdegir> #info so we save a lot of disk ios etc + cpu load. so that we can run 6 virtual deployment on one node without performance issues
14:32:31 <jmorgan1> s/other/of
14:32:43 <NikoHermannsEric> fdegir: ok Iwill use info now :-D
14:32:51 <fdegir> jmorgan1: https://wiki.opnfv.org/display/INF/Cross+Community+CI
14:32:52 <fdegir> :P
14:32:58 <fdegir> you have a nice diagram there
14:33:04 <NikoHermannsEric> #info documentation is not much there yet
14:33:20 <NikoHermannsEric> #info documentation will come until next week
14:33:24 <NikoHermannsEric> where to put doc?
14:33:37 <fdegir> NikoHermannsEric: https://wiki.opnfv.org/display/INF/ODL+3rd+Party+CI
14:33:50 <NikoHermannsEric> ok
14:34:00 <fdegir> let me just put our branding there
14:34:03 <fdegir> it is not 3rd party ci
14:34:05 <fdegir> XCI
14:34:36 <fdegir> https://wiki.opnfv.org/display/INF/ODL+XCI
14:34:50 <jmorgan1> haha... just got it X is cross
14:34:53 <fdegir> NikoHermannsEric: should we move on
14:34:57 <NikoHermannsEric> sure
14:35:02 <fdegir> jmorgan1: you are brilliant!
14:35:11 <NikoHermannsEric> fdegir: one more thing
14:35:14 <NikoHermannsEric> sorry
14:35:15 <fdegir> yes
14:36:11 <NikoHermannsEric> a request came up from functest to do patchset verification on same setup. Jose was asked from people from Huawei.
14:36:41 <fdegir> NikoHermannsEric: yes, I had a chat with Jose about that
14:36:45 <NikoHermannsEric> I know it is too early but at least I made advertisment for our idea. I think functest group liked it very much
14:36:50 <fdegir> NikoHermannsEric: we should strive for that
14:36:57 <NikoHermannsEric> ah ok. I am not working on that right now.
14:37:11 <fdegir> NikoHermannsEric: and we should try to reuse whatever you do for ODL XCI for functest and other projects too
14:37:20 <fdegir> NikoHermannsEric: it kind of falls under CI Evolution
14:37:36 <fdegir> to increase the CI Coverage across OPNFV projects
14:37:53 <fdegir> moving on
14:37:58 <fdegir> #topic AOB
14:38:00 <NikoHermannsEric> I told him that our resources are not full used right now and we could reuse even the ericsson-virtual pods for that
14:38:10 <NikoHermannsEric> ok
14:38:17 <fdegir> I would like to congratulate hwoarang for being a Releng committer
14:38:23 <fdegir> we expect moarrr
14:38:25 <jmorgan1> I finally got some systems. In the process of doing install and will take a look at XCI next week
14:38:29 <hwoarang> moarrr
14:38:42 <fdegir> #info hwoarang has been voted as a new committer to releng
14:38:59 <fdegir> #info jmorgan1 will start working on XCI soon
14:39:13 <fdegir> and we have someone with us today who might join us soon!
14:39:18 <fdegir> isn't it right mbuil?
14:39:45 <fdegir> he knows odl/sfc stuff and I believe he can help us a lot with stuff
14:40:02 <mbuil> :) maybe... but I find what you are doing very interesting, so I am currently your fan
14:40:37 <fdegir> mbuil: get in the queue for autograph
14:40:54 <fdegir> I welcome you in advance
14:41:00 <mbuil> thx
14:41:15 <fdegir> any last minute additions from anyone?
14:41:53 <fdegir> then thank you all for joining and meet you again in 2 weeks
14:41:58 <fdegir> #endmeeting