16:00:09 <frankbrockners> #startmeeting FDS synch up
16:00:09 <collabot> Meeting started Thu Oct 27 16:00:09 2016 UTC.  The chair is frankbrockners. Information about MeetBot at http://wiki.debian.org/MeetBot.
16:00:09 <collabot> Useful Commands: #action #agreed #help #info #idea #link #topic.
16:00:09 <collabot> The meeting name has been set to 'fds_synch_up'
16:00:17 <jlinkes> #info Juraj Linkes
16:00:19 <frankbrockners> roll call
16:00:20 <mmarsale1> #info Maros Marsalek
16:00:24 <frankbrockners> #info Frank Brockners
16:01:47 <raymondmaika> #info Raymond Maika
16:02:36 <frankbrockners> #info draft agenda: discuss key open items: (a) Colorado 2.0 update (b) vhost-user reconnect issue & follow-on problems (c) HA deploy & test  status on CENGN & Cisco FDS POD (d) hostconfig implementation (e) security groups and ODL cluster testing
16:03:53 <frankbrockners> #info Brief update on Colorado 2.0: David McBride is driving a discussion to move the Colorado 2.0 release date to Nov/10. This means we still have an opportunity to get our new fdio scenarios into Colorado 2.0.
16:04:53 <frankbrockners> #info For this to happen, we'd need to have the 4 test runs in around Nov/4 - and obviously the TSC has to approve the new release date.
16:05:22 <frankbrockners> jlinkes - could you give us an update on vhost-user reconnect and the related issues?
16:05:41 <jlinkes> #info we identified correct qemu and we need to use VPP 16.12 along with it
16:06:12 <jlinkes> #info but 16.12 vpp crashes when we recreate interfaces where traffic is flowing
16:06:28 <jlinkes> #info vpp guys are working on fixing that today - https://jira.fd.io/browse/VPP-516
16:06:45 <jlinkes> #info after this is fixed we're anticipating a full resolution of the issue
16:07:40 <jlinkes> that's all
16:07:57 <frankbrockners> thanks jlinkes - keeping fingers crossed
16:08:23 <frankbrockners> marcello__sestak - could you give us an update on HA deploy and testing?
16:08:39 <jlinkes> marcel is out on a teambuilding thing
16:08:40 <frankbrockners> raymondmaika - feel free to chime in as well
16:09:04 <jlinkes> as are the rest of the pantheon guys
16:09:24 <jlinkes> I don't like bowling, so I stayed at work during the bowling part
16:09:27 <frankbrockners> jlinkes - thanks for letting us know
16:09:43 <frankbrockners> jlinkes  - thanks for staying with us :-)
16:09:57 <raymondmaika> #info had an issue with our infra on Monday afternoon - while that was going I swapped out our nodes as I mentioned in the meeting on Monday
16:10:30 <raymondmaika> #info redeployed HA and lab was back up on Monday night, Marcel was running functest on the stack
16:10:45 <raymondmaika> jlinkes: knows more about results of those attempts than I do
16:10:53 <jlinkes> #info functest passed once on CENGN and since then we had problems getting past healthcheck
16:11:27 <frankbrockners> #info Marcel's functest run was succhessful once - results got submitted to testresults DB  (http://testresults.opnfv.org/reporting/functest/release/colorado/index-status-apex.html)
16:11:53 <frankbrockners> jlinkes - do you know what the status is on the Cisco FDS POD for HA deploy=
16:11:55 <jlinkes> we were seeing that vms were not getting ips from dhcp, I'm not sure about other problems
16:12:23 <raymondmaika> jlinkes: I will do some checking this afternoon for the functest issues
16:12:29 <jlinkes> #info the deployment on the FDS UCS-B pod finished successfully
16:12:45 <frankbrockners> jlinkes - did you run functest there?
16:13:10 <jlinkes> #info after configuring gallera, I was able to spin up two vms manually and check that they were pinging
16:13:44 <frankbrockners> ah - that reminds me of one question:
16:14:03 <jlinkes> #info there are problems with accessing vm consoles and ssh'ing into the vms - I had to connect through dhcp namespace
16:14:41 <frankbrockners> jlinkes - what is the patch for APEX that is pending to be merged which would avoid the gallera config changes?
16:14:46 <jlinkes> I tried troubleshooting the vm console access issue with trozet and so for I've found out that there are some problems with memcached
16:15:56 <jlinkes> frankbrockners: https://jira.opnfv.org/browse/APEX-337 and trozet mentioned that he's working on the gallera config changes because other scenarios might need that config
16:16:43 <jlinkes> frankbrockners: and I didn't run functest since I couldn't connect to the vms' floating ips
16:16:58 <frankbrockners> #info (per jlinkes) https://jira.opnfv.org/browse/APEX-337 and trozet mentioned that he's working on the gallera config changes because other scenarios might need that config
16:17:07 <frankbrockners> thanks jlinkes
16:17:43 <frankbrockners> does anyone have any updates on hostconfig/secgroups/ODL-clustering?
16:19:07 <frankbrockners> in addition - seanatcisco - are you around? if so - how far off are we from having a the ml2-vpp (i.e. os-nosdn-fdio-noha) from deploying and running functest? per my note above - we might be able to get it into the moved out Colorado 2.0
16:19:40 <jlinkes> Tomas mentioned that he's hit a new issue - "unknown vif type vhostuser"
16:19:56 <jlinkes> and he contacted some guys from Openstack
16:20:12 <jlinkes> that's all I know
16:20:30 <frankbrockners> #info on hostconfig development: Tomas mentioned that he's hit a new issue - "unknown vif type vhostuser"; Tomas contacted OpenStack folks
16:20:46 <frankbrockners> seanatcisco - r u here?
16:21:14 <frankbrockners> does not look like he's around
16:21:22 <frankbrockners> any other topics?
16:23:00 <jlinkes> don't think so
16:23:23 <frankbrockners> in case that isn't the case, am keeping fingers crossed that we (a) get the patch that fixes vhost-user reconnect ultimately (b) have raymond find the way to get HA to redeploy on CENGN (c) jlinkes/trozet find the issues with memcached and (d) tomas to get past the vif type issue...
16:23:29 <frankbrockners> thanks folks!
16:23:33 <frankbrockners> #endmeeting