16:59:29 <frankbrockners> #startmeeting FDS synch
16:59:29 <collabot`> Meeting started Thu Nov 10 16:59:29 2016 UTC.  The chair is frankbrockners. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:59:29 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:59:29 <collabot`> The meeting name has been set to 'fds_synch'
16:59:35 <frankbrockners> #info Frank Brockners
16:59:40 <mmarsale> #info Maros Marsalek
16:59:42 <jlinkes> #info Juraj Linkes
17:00:11 <frankbrockners> #info Agenda for today https://wiki.opnfv.org/display/meetings/FastDataStacks#FastDataStacks-Thursday,November10,2016
17:00:46 <michal-cmarada|2> #info Michal Cmarada
17:01:30 <frankbrockners> #info Agenda in a nutshell: (a) security groups testing / qemu issues (b) HA deployment (Cisco FDS POD and CENGN supermicro POD) - status on APEX-342, APEX-343, APEX-344 (c) hostconfig implementation / testing (d) L3 scenario - manual setup status (e) os-nosdn-fdio-noha test status
17:01:39 <tomas_c> #info Tomas Cechvala
17:02:26 <frankbrockners> jlinkes could you provide an updated on (a) qemu issues / sec group testing
17:02:53 <jlinkes> I tried the newest vpp today, which is 17.01
17:03:17 <jlinkes> it works worse than the previous version - the first application of security group to a vm crashes vpp
17:03:52 <jlinkes> whereas the previous version didn't crash and only prevented vm from communicating, and only after second application of a sec group
17:04:24 <frankbrockners> jlinkes - can you detail that crash - with api tracing etc. - ideally also with VPP built with debug on?
17:04:40 <jlinkes> I've sent an e-mail with api traces
17:05:00 <frankbrockners> do we things in an associated Jira ticket?
17:05:01 <jlinkes> I could try building vpp with debugs on and reproduce the issue
17:05:15 <jlinkes> Not yet
17:05:26 <frankbrockners> just trying to make sure that we follow the right process
17:05:38 <frankbrockners> so would be good if you could create a ticket
17:05:44 <frankbrockners> thanks
17:05:49 <jlinkes> fds ticket or vpp ticket? or both?
17:05:56 <frankbrockners> in vpp
17:06:04 <frankbrockners> and then a replica in fds
17:07:07 <frankbrockners> ok - let's move to HA deployments
17:07:21 <frankbrockners> jlinkes - that is you again...
17:08:10 <frankbrockners> any updates on either Cisco FDS POD or CENGN Supermicro POD?
17:08:23 <jlinkes> I tried to deploy ovs with performance options on the fds pod, unsuccessfully
17:08:35 <jlinkes> trozet haven't had time yet to look at the logs from the failure
17:09:46 <jlinkes> cengn - it seems that newer version of odl intoduced some issues - sometimes, vhostuser port is attempted to be attached to a bd before it's fully recreated
17:10:22 <jlinkes> michal-cmarada|2 and I think it might be a race condition, but it's not clear from logs
17:10:48 <jlinkes> we talked about strengthening logging in odl to better undestand if something is a race condition
17:10:55 <frankbrockners> does that mean that we can build ODL again?
17:11:10 <jlinkes> I'm in the process of trying out an older version of odl to see if it happens there too
17:11:19 <frankbrockners> ok - thanks jlinkes
17:11:30 <jlinkes> the odl version is from 5th Oct
17:11:49 <frankbrockners> ok
17:12:00 <frankbrockners> let's move to (c) hostconfig status
17:12:06 <frankbrockners> tomas_c - any updates?
17:12:16 <tomas_c> frankbrockners: yes, there is a patch on gerrit
17:12:26 <tomas_c> it's been tested couple of times against devstack
17:12:36 <tomas_c> and now i'm about to move to lab
17:12:43 <tomas_c> and test it there tomorrow morning
17:12:48 <tomas_c> i've been writing some UT
17:12:55 <tomas_c> so haven't spared some time for that
17:13:03 <tomas_c> so let's call it tomorrow morning
17:13:03 <frankbrockners> sounds good - thanks tomas_c
17:13:09 <tomas_c> yw
17:13:38 <frankbrockners> ok - (d) L3 implementation is next
17:13:45 <frankbrockners> michal-cmarada|2 - any updates?
17:13:52 <michal-cmarada|2> actualy yes
17:14:23 <frankbrockners> did you get arp to work=
17:14:25 <michal-cmarada|2> I managed to get the demo running again thanks to some dependencies from tomas and vlado
17:14:25 <frankbrockners> ?
17:14:37 <frankbrockners> michal-cmarada|2 - great news
17:14:57 <michal-cmarada|2> and when ARP is set staticly I can ping an external node
17:15:00 <frankbrockners> let's plan to show it in next Monday's synch meeting
17:15:48 <michal-cmarada|2> I wrote an email to Ole to figure out if there is a way to setup vpp to automaticaly respond to ARP requests on NAT IPs
17:16:03 <frankbrockners> next up is (e) os-nosdn-fdio-noha test status
17:17:06 <frankbrockners> seanatci_ - any updates? I know that you got things to work, though I don't see results in the testresults DB yet
17:17:41 <frankbrockners> which means that there are also still some issues when deployed as part of the OPNFV CI/CD
17:17:50 <frankbrockners> seanatcisco - r u around?
17:18:38 <jlinkes> cisco jabber says he's in a meeting
17:18:58 <frankbrockners> jlinkes - ok - let's get updates from him offline
17:19:11 <frankbrockners> is there anything else to discuss here / today?
17:19:13 <jlinkes> frankbrockners: ticket for vpp crash: https://jira.fd.io/browse/VPP-528
17:19:20 <frankbrockners> thanks jlinkes
17:19:37 <jlinkes> I also found a HC bug today
17:19:40 <jlinkes> let me find it
17:20:03 <mmarsale> that's the one https://jira.fd.io/browse/HONEYCOMB-292
17:20:23 <jlinkes> https://jira.fd.io/browse/HONEYCOMB-292
17:20:28 <jlinkes> ah too slow
17:20:40 <mmarsale> after some time requests times out between VPP and HC ... currently executing a long running test to reproduce ... not sure where;s the problem just yet
17:21:02 <mmarsale> but will keep looking into it
17:21:09 <frankbrockners> thanks - jlinkes - could you reply to your own email about the qemu issue and mention VPP-528? thanks
17:21:42 <jlinkes> sure frankbrockners
17:22:46 <frankbrockners> thanks jlinkes
17:22:58 <frankbrockners> sounds like we're done for today
17:23:01 <frankbrockners> anything else?
17:23:36 <jlinkes> I'm done
17:23:47 <tomas_c> tomas_c: me too
17:23:51 <frankbrockners> ok - thanks everyone - have a good day / evening
17:23:54 <frankbrockners> #endmeeting