15:59:25 <frankbrockners> #startmeeting FDS synch
15:59:25 <collabot> Meeting started Thu Jul 14 15:59:25 2016 UTC.  The chair is frankbrockners. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:59:25 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic.
15:59:25 <collabot> The meeting name has been set to 'fds_synch'
15:59:35 <frankbrockners> #info Frank Brockners
15:59:47 <frankbrockners> hey folks - could those of you who are around #info in?
16:02:05 * frankbrockners feels lonely....
16:02:27 <michal-cmarada> #info Michal Cmarada
16:02:27 <jlinkes_> #info Juraj Linkes
16:04:06 <frankbrockners> ok... let's get moving and see who else chimes in...
16:04:30 <frankbrockners> could anyone (Juraj) give a quick status on what works and what does not on CENGN/Paris labs?
16:05:25 <jlinkes_> #info CENGN - the deployment ends up with no keystone services and keystone endpoints configured, we don't know why this is happening yet
16:06:15 <jlinkes_> #info CENGN - next steps would be configuring one service manually and seeing if that works
16:06:35 <trozet> jlinkes_: we have seen this with ovs dpdk
16:07:06 <frankbrockners> trozet - any workaround?
16:07:06 <trozet> jlinkes_: has something to do with the VIP/VRRP
16:07:28 <trozet> frankbrockners: yeah don't muck with the controller
16:07:38 <trozet> frankbrockners: let it have regular OVS on
16:07:49 <trozet> frankbrockners: dont set netdev in neutorn to True either
16:07:54 <jlinkes_> we're deploying with regular ovs
16:08:09 <trozet> jlinkes_: can you verify there is a VIP on br-ex by doing ip addr show
16:08:31 <trozet> jlinkes_: also fpan just fixed a bug where the control plane VIP was being placed on br-ex intead of the ctlplane interface
16:09:11 <jlinkes_> trozet: only public vip should be on br-ex?
16:09:19 <trozet> jlinkes_: yeah
16:09:55 <jlinkes_> trozet: there was the public vip on br-ex, but the public network was not configured properly
16:10:25 <trozet> jlinkes_: ok, what happens is the undercloud tries to create the endpoints post install
16:10:31 <jlinkes_> #info Paris - not sure why the deployment works, will need to look at nic ordering in more detail, that's where I'm suspecting the root cause it atm
16:10:35 <frankbrockners> jlinkes_ - what was that issue with the port conflicts for keystone
16:10:52 <trozet> jlinkes_: why is the public network not correct?
16:11:21 <jlinkes_> trozet: it needs to be in a vlan, but apex doesn't work with mix of vlan/nonvlan networks
16:12:01 <jlinkes_> frankbrockers: we're not 100% sure, but what probably happened is that a reboot of the control node screwed everything up
16:12:16 <trozet> jlinkes_: can you be more specific to what doesn't work when you mix vlan/non vlan
16:12:24 <trozet> jlinkes_: that was just resolved 3 weeks ago to add vlan support
16:12:33 <trozet> jlinkes_: and radez just fixed a bug last night where public vlan was wrong
16:12:47 <frankbrockners> jlinkes_ - could you set aside some time today with Tim and Sean to discuss the current set of issues in detail? Tim is back from PTO and can help
16:12:49 <trozet> https://gerrit.opnfv.org/gerrit/#/c/16787/
16:13:12 <jlinkes_> frankbrockners: sure
16:13:42 <frankbrockners> thanks jlinkes_
16:13:43 <jlinkes_> trozet: three networks were configured as native and the public one was configured as vlan 1199, but it wasn't configured that way
16:14:17 <trozet> jlinkes_: that commit I just posted would impact that type of deployment
16:14:23 <jlinkes_> trozet: eh, it was configured that way in network_settings.yaml, but on the actual deployment that wasn't the case
16:14:34 <jlinkes_> trozet: ok, will try with that
16:14:41 <trozet> jlinkes_: i can try to deploy  what you just mentioned on my lab and see what I get
16:16:06 <jlinkes_> trozet: the fix should be in the latest rpms, right?
16:17:10 <trozet> checking
16:17:30 <jlinkes_> #info will try to redeploy CENGN with latest rpm which should fix the issue of mixing public vlan with nonvlan networks and then debug issues with tim if any come up
16:17:58 <trozet> jlinkes_: yes
16:21:47 <frankbrockners> jlinkes_ - could you also synch with Tim on the keystone (and eventual port binding conflicts) that you mentioned yesterday?
16:23:59 <frankbrockners> anything else to be surfaced for today - or are we done?
16:24:04 <jlinkes_> frankbrockners: sure
16:24:11 <jlinkes_> I'm done
16:24:26 <trozet> jlinkes_: i need to run for about an hour, send me anything you have and I can try to reproduce on my setup
16:24:27 <frankbrockners> martin_sunal - anything from your end?
16:25:05 <frankbrockners> looks like martin_sunal isn't around
16:25:09 <frankbrockners> let's call it a day
16:25:12 <frankbrockners> thanks everyone
16:25:15 <frankbrockners> #endmeeting