16:01:13 <Woj> #startmeeting FDS-weekly
16:01:13 <collabot> Meeting started Thu Jul 28 16:01:13 2016 UTC.  The chair is Woj. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:01:13 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:01:13 <collabot> The meeting name has been set to 'fds_weekly'
16:01:49 <Woj> Hi everyone... Starting meeting in Frank's absence
16:02:04 <Woj> #topic Agenda
16:02:09 <tomas_c> tomas_c: Hello
16:04:23 <Woj> 1. Open-issues: https://wiki.opnfv.org/display/fds/FDS+Open+Issues
16:04:43 <Woj> 2. Any other business
16:05:18 <Woj> Any other points to discuss?
16:05:40 <Woj> just bring them up... in the meantime let's go with 1.
16:06:02 <Woj> #topic Open Issues
16:06:47 <Woj> A number of jira tickets concerning general apex install issues
16:07:20 <Woj> Tim, any progress on the VPP/huge-page settings?
16:08:37 <Woj> Guess Tim's not here
16:09:18 <Woj> #info No update on open jira issues (for lack of attendees)
16:09:45 <Woj> #topic ucs-c update
16:10:19 <Woj> @michal-cmarada ?
16:10:19 <collabot> Woj: Error: "michal-cmarada" is not a valid command.
16:10:31 <michal-cmarada> ok I can start
16:11:37 <michal-cmarada> For now we have som minor issues with sub interfaces creation
16:11:50 <michal-cmarada> but we are working on that right now
16:12:29 <Woj> any estimate when we'll be clear with that?
16:12:33 <tomas_c> Woj: we are also waiting for more details about the switch connected to compute nodes
16:12:40 <michal-cmarada> we also had problems with som changes in v3po model but this was fixed today and patch was merged today
16:13:12 <michal-cmarada> I think that we can fix the problem regarding subinterfaces tommorow
16:13:25 <Woj> Anything you need to progress?
16:13:51 <tomas_c> not yet
16:13:58 <tomas_c> we just need to know the env entirely
16:14:01 <michal-cmarada> not, now. we are just debuging this localy and then we will try it in the lab
16:15:01 <tomas_c> we hope we will know where we are in the next few days
16:15:24 <tomas_c> tomorrow or mon tue
16:15:38 <michal-cmarada> hopefully the VLAN case wil be ready next week
16:16:02 <Woj> Ok. Let's please try to make it tomorrow :-)
16:16:07 <michal-cmarada> at least the tests in UCS-C
16:16:21 <michal-cmarada> we will try
16:16:23 <Woj> Have you tried VxLAN setup there ?
16:16:52 <michal-cmarada> i think that martin did that, but I am not sure
16:17:27 <tomas_c> we focus on VLAN now
16:17:36 <tomas_c> but vxlan had worked before
16:17:57 <Woj> Ok, so on POD2 - UCS-B, VxLAN has not been working (tunnels were failing to get created)
16:18:02 <michal-cmarada> VXLAN was also tested on POD2 i think
16:18:11 <Woj> this week
16:18:36 <tomas_c> ok, we will try to test both cases asap
16:18:43 <Woj> would be great if you can check out the VxLAN in your setup too - should be straighforward
16:18:44 <tomas_c> we are on it at the moment
16:18:47 <Woj> tx
16:18:48 <michal-cmarada> so for this week i dont know, but there was the issue with v3po models that was fixed today
16:19:04 <Woj> Moving on...
16:19:28 <Woj> #topic Pod2 update
16:20:27 <Woj> vNIC issue regarding corrupted IP packets is supposedly fixed. Have been unable to verify though.
16:22:16 <Woj> Not strictly pod2 related, but verified there: I switched VPP to act as vhostuser server. This is desirable to give quiucker failure notification in nova if things go wrong.
16:22:54 <Woj> It worked fine, but needed changes to permissions & SElinux settings. Opened jira tickets for these
16:23:37 <Woj> ODL patch is https://git.opendaylight.org/gerrit/#/c/42620/
16:24:02 <Woj> corresponding neutron-odl patch is still needed
16:24:23 <Woj> #topic L3 design
16:25:49 <tomas_c> l3 design within tenant networks is completed
16:26:08 <tomas_c> any comments are apprecitated though
16:26:32 <tomas_c> external networking needs a bit more discussion in terms of topology design
16:26:42 <tomas_c> how to route outside
16:26:52 <Woj> Had a useful call with apex team today to clear out things. Arriving at L3 VPP gateway for Colorado looks doable. Needed are: 1. Triple0 needs to setup VPP on control node 2. ODL L3+sNAT config
16:27:15 <Woj> 3. (as tomas_c writes) we need to discuss in some more detail the neutron info mapping
16:27:50 <tomas_c> Woj: sure
16:27:53 <Woj> There is a possibility to fallback to using the regular L3-agent should we get stuck on the above. Woj to investigate
16:28:55 <Woj> idea is to hookup L3 agent via a tap interface port.
16:29:11 <Woj> into vpp running on the control node
16:29:32 <Woj> #topic any other stuff?
16:30:13 <Woj> Anyone blocked on anything not covered?
16:31:20 <Woj> Seems not, so I'll be closing the meeting
16:31:23 <Woj> thanks all
16:31:54 <Woj> #endmeeting