14:00:15 <fdegir> #startmeeting Cross Community CI
14:00:15 <collabot`> Meeting started Wed Jan 31 14:00:15 2018 UTC.  The chair is fdegir. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:15 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic.
14:00:15 <collabot`> The meeting name has been set to 'cross_community_ci'
14:00:23 <fdegir> hello everyone
14:00:33 <fdegir> as hwoarang says, it is that time of the week again
14:00:39 <fdegir> #topic Rollcall
14:00:39 <hw_wutianwei> hello
14:00:52 <fdegir> hi hw_wutianwei
14:00:55 <ttallgren> #info Tapio Tallgren
14:00:59 <mbuil> #info Manuel Buil
14:01:03 <hw_wutianwei> #info Tianwei Wu
14:01:06 <joekidder> #info Joe Kidder
14:01:34 <fandrieu> #info Fabien Andrieux
14:01:40 <david_Orange> #info David Blaisonneau
14:01:46 <fdegir> let's start with the first topic and others can join on the way
14:02:05 <fdegir> agenda was on https://etherpad.opnfv.org/p/xci-meetings
14:02:13 <fdegir> but etherpad seems to be down
14:02:22 <fdegir> anyone else having trouble opening it?
14:02:26 <hwoarang> #info Markos Chandras
14:02:45 <joekidder> fdegir: yes no etherpad
14:02:51 <hw_wutianwei> I can't open
14:02:52 <fdegir> aricg: bramwelt: ^
14:03:14 <fdegir> it was pretty similar to previous week's agenda
14:03:22 <fdegir> and the first topic is
14:03:23 <ttallgren> "502 Bad Gateway"
14:03:26 <fdegir> #topic Functest Healthcheck Status
14:03:41 <fdegir> some of you might know that, we run CI within VMs
14:03:55 <mardim> #info Dimitrios Markou
14:04:04 <peri> #info Periyasamy Palanisamy
14:04:07 <fdegir> meaning that all the VMs (opnfv, controller, compute) get created in distro VMs to ensure we always have clean environment
14:04:14 <fdegir> and increase the no of resources we have
14:04:34 <fdegir> but we are seeing strage failures with functest healthcheck when running things this way
14:05:12 <fdegir> to be more precise, the snaps testcase test_add_remove_volume test fails attaching/detaching volume from instance
14:05:30 <fdegir> we have been trying to find the root cause but failed to do so
14:05:56 <fdegir> because of this, we haven't been able to bump shas and will not do that until we either find the problem and fix it
14:06:06 <fdegir> or switch back to running things on the host directly
14:06:13 <mbuil> fdegir: the volume does not attach to the instance?
14:06:24 <hw_wutianwei> if using ceph, it maybe pass this failures about volume
14:06:25 <david_Orange> fdegir, is it working on a stable branch ?
14:06:30 <fdegir> which results in reduced no of CI resources and more importantly removes the ability to run things in always clean environment
14:06:56 <fdegir> mbuil: when I tried, I was able to attach but unable to remove the volume
14:07:07 <mbuil> ok
14:07:10 <fdegir> hw_wutianwei: can you give some more details please?
14:07:43 <fdegir> david_Orange: it works on latest but not when we run things within vm
14:08:02 <hw_wutianwei> these testcase can be passed in compass4nfv.
14:08:25 <fdegir> david_Orange: when we have nested virtualization
14:08:30 <hwoarang> we are not using ceph though
14:08:41 <david_Orange> fdegir, ok
14:09:21 <hwoarang> fwiw here is the traceback from detaching the volume
14:09:23 <hwoarang> http://paste.opensuse.org/27139788
14:09:35 <hwoarang> from the nova-compute.log on compute node
14:10:03 <hw_wutianwei> In my opinion, it is related to storage.
14:10:13 <hwoarang> it's  not so much with detaching the volume but with neutron but i am still trying to decipher things
14:10:58 <fdegir> so, long story short, we will continue looking into this 1 or 2 more days
14:11:13 <electrocucaracha> #info Victor Morales
14:11:15 <fdegir> and if we can't come to a solution, we will change how we run things in CI and stop using VMs
14:11:22 <ttallgren> Storage tends to be a bottleneck. Do you see high storage load in the virtual host?
14:11:29 <fdegir> which in turn will result longer queue times for patch verification
14:11:51 <mbuil> libvirtError... seems like a nested issue
14:12:13 <fdegir> when I create an instance manually, I am able to attach a volume
14:12:20 <fdegir> but I fail to remove the volume
14:12:29 <fdegir> and I have different error comparing to hwoarang
14:13:05 <fdegir> so this was all about this topic
14:13:29 <fdegir> if anyone is willing to give a try, just ping hwoarang and me and we can summarize how you can reproduce the issue
14:13:53 <fdegir> moving to the next high prio topic
14:13:57 <fdegir> #topic Baremetal Status
14:14:28 <fdegir> testing is one thing to enable patchset verification and scenario promotion post merge jobs
14:14:46 <fdegir> and baremetal is the next one that should happen once a scenario gets promoted
14:15:24 <david_Orange> fdegir, in my side i am still working on it
14:15:31 <fdegir> mbuil is going to start working on this for baremetal
14:15:42 <fdegir> with the help from david_Orange :)
14:15:49 <david_Orange> fdegir, actually on the DC to debug a pxe issue
14:16:07 <mbuil> david_Orange: you mean you still need to update your patches?
14:16:40 <david_Orange> mbuil, yes, sorry i had a long week and can not focus on that
14:17:35 <david_Orange> mbuil, fdegir is this baremetal will use pdf/isd as source or use the actual fixed ip config .
14:17:36 <david_Orange> ?
14:17:43 <mbuil> david_Orange: ok, no worries :). Should I wait then?
14:17:45 <fdegir> david_Orange: pdf/idf
14:18:16 <david_Orange> mbuil, yes, please, i will keep you in touch when we will test the baremetal infra deployment
14:18:31 <david_Orange> the focus on OSA
14:18:37 <david_Orange> then focus on OSA
14:19:03 <fdegir> thanks mbuil david_Orange
14:19:16 <fdegir> if someone else wants to work on it with a different scenario as well (such as os-nosdn-nofeature, etc.), just ping me and I can find a pod
14:19:50 <fdegir> but that person needs to wait for david_Orange as well unless he or she prefers going to hard way and doing things from scratch
14:20:14 <fdegir> moving to the usual topics
14:20:22 <david_Orange> i really hope i can patch 'till the end of week
14:20:28 <fdegir> #topic Scenario/Feature Status: os-odl-sfc
14:20:38 <fdegir> mbuil: mardim: any update for sfc?
14:21:41 <mbuil> fdegir: not really. All bugs which appeared when moving to the new SHAs are fixed
14:22:14 <mbuil> We are ready to make that move, although we would prefer to move to newer SHAs where those fixes are included. Apart from that, I have nothing
14:22:58 <fdegir> mbuil: we will hopefully move to the new SHAs by next week
14:23:20 <fdegir> thx mbuil
14:23:24 <fdegir> #topic Scenario/Feature Status: os-odl-bgpvpn
14:23:30 <mardim> fdegir, nothing from my side I am occupied with Tacker problems
14:23:39 <fdegir> peri: I've seen the blueprint got merged
14:23:45 <fdegir> peri: anything more?
14:24:00 <peri> yes, I have tested this scenario with ODL nitrogen
14:24:26 <peri> but looks like some issue with this version. raised a bug in netvirt https://jira.opendaylight.org/browse/NETVIRT-1071
14:25:16 <peri> but it works fine with ODL version (i.e BGP peering). so i request you to review the upstream patches
14:25:32 <peri> i meant ODL carbon version
14:25:39 <fdegir> ok
14:25:59 <fdegir> so there is progress which is good and at the same time a bug is identified
14:26:50 <peri> yes, but we have get upstream reviews to be reviewed and merged
14:26:57 <peri> its long pending
14:27:16 <peri> https://review.openstack.org/#/c/522598/ and https://review.openstack.org/#/c/523907/
14:28:07 <hwoarang> sorry these went under the radar because of zuul -1
14:28:14 <peri> gate job /c/522598 is failing continuously. need someone help to figure out the reason
14:28:35 <fdegir> as hwoarang says, zuul seems to be pretty shaky lately
14:28:54 <fdegir> I've been following zuul status and there is at least 1 issue every day
14:29:19 <peri> but it works for the review https://review.openstack.org/#/c/538933/
14:29:22 <hwoarang> peri: zuul + xenial was massively broken yesterday
14:29:28 <mbuil> peri: try again. That Ubuntu gate was failing for me but yesterday evening it worked
14:29:30 <hwoarang> please recheck both
14:29:38 <fdegir> if you are using twitter, you can follow the news :)
14:29:38 <fdegir> https://twitter.com/openstackinfra?lang=en
14:30:05 <fdegir> peri: I suppose that's all
14:30:20 <peri> yes, thats all :)
14:30:21 <mbuil> fdegir: we should use Twitter too
14:30:25 <fdegir> thanks peri
14:30:32 <fdegir> mbuil: why not :)
14:30:39 <hwoarang> a bot to post openstackinfra tweets here :)
14:30:39 <fdegir> but we first need to test things, you know
14:31:28 <fdegir> we have our own spambot
14:31:31 <fdegir> OPNFV-Gerrit-Bot: ping?
14:31:37 <fdegir> moving
14:31:43 <fdegir> #topic Scenario/Feature Status: k8-nosdn-nofeature
14:31:59 <fdegir> hw_wutianwei: I see you've been sending few more patches lately
14:32:05 <hw_wutianwei> #link  https://gerrit.opnfv.org/gerrit/#/c/50213/
14:32:26 <hw_wutianwei> fdegir: yep
14:32:49 <hw_wutianwei> it finished and passed the CI verify. I think it can be megered.
14:32:56 <fdegir> hw_wutianwei: +1
14:33:08 <hw_wutianwei> thanks the people review this patch and feedback, especially hwoarang
14:33:23 <hw_wutianwei> and mbuil
14:33:37 <hwoarang> yep we can get it in
14:33:46 <fdegir> hw_wutianwei: what would be the next step?
14:33:58 <fdegir> hw_wutianwei: opensuse/centos support or ?
14:33:59 <hw_wutianwei> should we create kubernetes verify job?
14:34:07 <hw_wutianwei> fdegir: yep
14:34:10 <hwoarang> a xenial job would be nice
14:34:11 <fdegir> hw_wutianwei: that I can take care of
14:34:17 <hw_wutianwei> I will support centos
14:34:23 <fdegir> hw_wutianwei: I mean the job stuff
14:34:24 <hw_wutianwei> first
14:34:25 <hwoarang> i will do the suse bit
14:34:31 <fdegir> good
14:34:34 <hw_wutianwei> hwoarang: thanks
14:34:43 <fdegir> so please keep an eye on releng patches coming days
14:34:50 <fdegir> creating jobs for k8s
14:34:57 <fdegir> or adding it to existing jobs
14:35:23 <fdegir> thanks hw_wutianwei for taking this till the end
14:35:31 <hwoarang> yeah good job
14:35:41 <hw_wutianwei> one more thing, could you give me a centos vm?
14:35:49 <hw_wutianwei> fdegir: ^^
14:36:20 <fdegir> hw_wutianwei: we have one node that's not used by anyone but the OS installation failed on it
14:36:29 <fdegir> electrocucaracha: did you try reinstalling os on pod21-jump again?
14:36:52 <electrocucaracha> fdegir: nope, but I can try it today
14:37:04 <fdegir> electrocucaracha: that would be good
14:37:13 <fdegir> hw_wutianwei: we put ubuntu 16.04 on the nodes by default
14:37:21 <hw_wutianwei> fdegir, electrocucaracha:  It doesn't matter, I will try to find one
14:37:48 <fdegir> hw_wutianwei: we need to fix the node anyways so electrocucaracha can try agian and it works, you get that one
14:38:05 <hw_wutianwei> fdegir: ok, thank you
14:38:21 <fdegir> electrocucaracha: I'll ping you tomorrow if you don't ping me earlier
14:38:34 <electrocucaracha> fdegir: ok
14:38:36 <fdegir> moving to the next features
14:38:39 <fdegir> Taseer: around?
14:39:14 <fdegir> #topic Scenario/Feature Status: congress/blazar/masakari
14:39:30 <fdegir> I think the status for these are same as last week
14:39:40 <fdegir> the last patch for congress is still under review
14:40:02 <fdegir> blazar blueprint got accepted and Taseer working on role
14:40:24 <fdegir> masakari team hasn't started and will not start working on it until queens is out
14:41:01 <fdegir> fandrieu: should we talk about vpp?
14:41:14 <fandrieu> yes
14:41:23 <fandrieu> can give you a quick status
14:41:24 <fdegir> #topic Scenario/Feature Status: vpp
14:41:29 <fdegir> please go ahead
14:41:53 <fandrieu> Almost there. vpp and networking_vpp agent deployed and configured.
14:42:05 <fandrieu> Still have to figure out a few plumbing issues.
14:42:21 <fandrieu> Also have a question for you guys
14:42:36 <fdegir> yes
14:43:04 <fandrieu> I created a os-nosdn-vpp scenario for that. I realized later that it would be os-nosdn-vpp scenario with vpp ml2 plugin
14:43:14 <fandrieu> What would be the right way to integrate ?
14:43:53 <fandrieu> I meant os-nosdn-nofeature with vpp
14:43:53 <fdegir> have you looked at this?
14:43:54 <fdegir> https://wiki.opnfv.org/pages/viewpage.action?pageId=12390152#OnboardingProjects/ScenariostoXCI-StructuringtheWorktoOnboardtoXCI
14:44:39 <fandrieu> Yes. Not clear to me though
14:44:49 <fdegir> fandrieu: ok, will come to there after a question
14:44:59 <fdegir> isn't this scenario os-nosdn-fdio ?
14:45:27 <fdegir> rather than os-nosdn-vpp or os-nosdn-nofeature
14:45:28 <fandrieu> might be. fdio is the project. vpp is the vswitch technology
14:45:53 <fdegir> I think vpp integration is done under the scenario os-nosdn-fdio
14:46:12 <fandrieu> OK. will move into it then
14:46:18 <fdegir> we can check that and if that's the case, we have examples for you which you can take a look at and see how it can be done
14:46:25 <fdegir> about what needs to be done
14:46:33 <fdegir> first, you need a blueprint in upstream osa
14:46:45 <fdegir> proposing vpp integration
14:47:01 <jmorgan1> #info Jack Morgan
14:47:03 <fdegir> and in parallel to that, you can start upstreaming ansible roles
14:47:23 <fdegir> and in opnfv, you do opnfv scenario stuff similar to sfc or bgpvpn
14:47:53 <fandrieu> ok. clearer now.
14:47:54 <fdegir> I'll send an email to you with examples and include others who went through this and can help you doing this
14:48:17 <fdegir> the crucial part in this is to put your opnfv scenario role in fds repo, not in xci repo
14:48:17 <fandrieu> yep. kept the blueprint in mind but did not write it yet
14:48:45 <fandrieu> That was my next question
14:49:00 <fandrieu> For the sake of development I worked in xci/scenarios till now
14:50:44 <fdegir> the scenario will end up in <fds repo>/scenarios/os-nosdn-fdio/role/os-nosdn-fdio folder
14:50:56 <fdegir> please check sfc in this file  https://gerrit.opnfv.org/gerrit/gitweb?p=releng-xci.git;a=blob;f=xci/opnfv-scenario-requirements.yml
14:51:13 <fandrieu> will do. thanks
14:51:47 <fdegir> and you can ask how the scenarios are plugged in to xci when you start looking into it
14:51:58 <fdegir> hwoarang is the mastermind behind the mechanism
14:52:19 <fdegir> thanks fandrieu and you will get a long mail from me
14:52:46 <fandrieu> fdegir: will try to respond with a long mail as well :)
14:52:57 <fdegir> please do :)
14:53:00 <fdegir> #topic Kolla Wars
14:53:07 <fdegir> david_Orange: electrocucaracha: anything to say?
14:53:20 <david_Orange> fdegir, this is not a war :D
14:53:34 <electrocucaracha> I have rebased the latest changes on the draft
14:53:50 <electrocucaracha> I think the aio flavor is working in ubuntu and centos
14:53:50 <fdegir> david_Orange: I like using the term, don't blame me
14:53:52 <fdegir> blame Ray
14:53:52 <jmorgan1> yes, changte topic to Kolla for the win! (FTW)
14:54:02 <david_Orange> fdegir, i know :)
14:54:22 <fdegir> jmorgan1: that's not as fun, sorry
14:54:31 <jmorgan1> agreed
14:54:40 <david_Orange> fdegir, in my side i started to write my view on pdf/idf integration (that will also impact next baremetal test)
14:54:56 <david_Orange> fdegir, i will send it soon
14:55:04 <fdegir> david_Orange electrocucaracha: before diving into details, can you come up with a short/brief info about this?
14:55:15 <fdegir> as asked last week
14:55:27 <fdegir> so everyone knows what the intentions are
14:55:49 <fdegir> from last week's meeting: david_Orange will come up with a high level overview/plan and present it to team to collect feedback
14:55:51 <david_Orange> as i said, i have prepared a short text to explain how we can do that and set some roles that can be reused
14:56:13 <fdegir> it wasn't just about pdf/idf
14:56:34 <electrocucaracha> fdegir: do you mean the main goal of that effort?
14:56:36 <fdegir> it was general like why kolla, what are the intentions and so on
14:56:39 <fdegir> electrocucaracha: yes
14:56:55 <david_Orange> no, but mainly about how pdf/idf can impact today steps
14:57:16 <david_Orange> fdegir, humm i did not get that :)
14:57:38 <fdegir> pdf/idf and other impacts are after talking about the goals with having kolla in xci
14:57:42 <electrocucaracha> fdegir: I think that main idea is support an additional OpenStack installer which offers a containerized way
14:58:10 <fdegir> electrocucaracha: david_Orange: we have 3 minutes left so if you can come up with this type of overview, we can talk about it next week
14:58:51 <fdegir> I take the silence as "yes, we can do that"
14:58:54 <david_Orange> fdegir, ok
14:59:02 <david_Orange> fdegir, SIR YES SIR !
14:59:07 <david_Orange> fdegir, is it better ?
14:59:08 <fdegir> david_Orange: :)
14:59:14 <electrocucaracha> fdegir: yes, I think that I can elaborate better next week
14:59:21 <fdegir> thanks
14:59:29 <fdegir> #topic AOB
14:59:33 <fdegir> before we end today's meeting
14:59:39 <fdegir> anyone wants to add anything?
14:59:43 <electrocucaracha> yes,
14:59:43 <david_Orange> electrocucaracha, please feel free to ping me to talk about that
14:59:57 <electrocucaracha> the patch that adds proxy support is ready for reviews
15:00:07 <ttallgren> I am working on offline mode for CentOS. Not sure what will be done with that
15:00:19 <fdegir> ttallgren: offline mode?
15:00:20 <electrocucaracha> This one https://gerrit.opnfv.org/gerrit/#/c/45383/
15:00:46 <ttallgren> Installing the XCI from locally installed repos (rpm + git)
15:00:56 <electrocucaracha> I'm using this values for non-proxy https://gist.github.com/electrocucaracha/3b83004b0233b99c480546d895939839#file-run_jenkins_test-sh
15:01:10 <fdegir> ttallgren: that's interesting cause this is one of the things many people asking to upstream osa
15:01:15 <fdegir> enabling offline deployments
15:01:21 <hwoarang> ttallgren: that needs work on upstream osa and there is a spec for it
15:01:32 <hwoarang> better talk to the author of the spec otherwise i smell duplication
15:01:42 <jmorgan1> ttallgren: do you mean having a local mirror (rpm/git)?
15:01:48 <fdegir> hwoarang: can you put the link if you have that?
15:01:54 <ttallgren> jmorgan1: Yes
15:02:32 <fdegir> before I end the meeting
15:02:40 <hwoarang> http://git.openstack.org/cgit/openstack/openstack-ansible-specs/tree/specs/queens/python-build-install-simplification.rst
15:03:02 <fdegir> an XCI update will be given to opnfv community tomorrow during weekly tech discussion meeting
15:03:02 <hwoarang> that's for wheel packages but it's still part of the offline installation
15:03:03 <fdegir> https://wiki.opnfv.org/display/PROJ/Weekly+Technical+Discussion
15:03:14 <hwoarang> offline is not just about rpms
15:03:15 <fdegir> please join if you have time
15:03:34 <fdegir> thank you all and talk to you next week
15:03:36 <fdegir> #endmeeting