13:06:23 <morgan_orange> #startmeeting integration weekly meeting 10/07/2019 13:06:23 <collabot`> Meeting started Wed Jul 10 13:06:23 2019 UTC. The chair is morgan_orange. Information about MeetBot at http://wiki.debian.org/MeetBot. 13:06:23 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic. 13:06:23 <collabot`> The meeting name has been set to 'integration_weekly_meeting_10_07_2019' 13:06:28 <morgan_orange> #topic participants 13:06:35 <morgan_orange> #info Yang 13:06:37 <morgan_orange> #info Brian 13:06:41 <morgan_orange> #info Daniel Rose 13:06:46 <morgan_orange> #info Gary 13:06:51 <morgan_orange> #info joey Zhou 13:07:08 <morgan_orange> #info Kezystof Kuzmicki 13:07:13 <morgan_orange> #info Lei Shi 13:07:16 <morgan_orange> #info Marco 13:07:22 <morgan_orange> #info Paul Ionu Vaduva 13:07:30 <morgan_orange> #topic Quick update 13:08:12 <morgan_orange> #info release officially announced 13:08:27 <morgan_orange> #info still some issues tracked in the daily "war room" meeting 13:09:02 <morgan_orange> #info different behavior detected on Orange lab (Dublin daily chain) 13:09:23 <morgan_orange> #info probelm at startup components are waiting other components, it seems that there are some race conditions 13:11:28 <morgan_orange> #info investigation on installation shall be done, helm charts shall be reviewed and components shall manage more properly their dependency 13:13:04 <morgan_orange> #info orchestration at installation is needed - timers are not enough 13:16:52 <morgan_orange> #info helm is maybe not enough for the the job 13:17:02 <morgan_orange> #topic El Alto integration tasks 13:17:22 <morgan_orange> #info priority 1 don't break the build (sanity daily run) 13:18:24 <morgan_orange> #info extending gating to component 13:23:01 <morgan_orange> #info risk reported to TSC and PTL ..everybody agree during DDF, no decisions taken so far 13:23:39 <morgan_orange> #info discussions on the gating on SO 13:24:23 <morgan_orange> #info Brian suggest we could give a try on one tenant of the windriver lab. SO shall be deployed on an ONAP using fixed versions for the other components 13:24:49 <morgan_orange> #info a subset of the components is needed (not all the modules) 13:27:52 <morgan_orange> #action Yang evaluate SO gating in windriver lab 13:28:30 <morgan_orange> #action morgan_orange xcheck with Sylvain on creating a chain for SO to windriver lab 13:29:35 <morgan_orange> #info SB00 installed with latest Dublin 13:31:42 <morgan_orange> #info last run of Dublin on Orange lab => healthcheck almost all OK but impossible to onboard/instantiate, investigation in progress 13:32:02 <morgan_orange> #info ONAP deployment with smaller footprint 13:32:25 <morgan_orange> #info Bell Canada tried to divide by 2 the resource usages => aciton on OOM team 13:33:04 <morgan_orange> #info on windriver lab, Yang suggest to setup quotas to avoid resources starvation few eeks before the release as already experienced 13:33:56 <morgan_orange> #info Yang suggests also to deploy a sharable ONAP to save resources 13:37:44 <morgan_orange> #info several projects could use shared components e.g. 1 AAF for N tenants 13:39:07 <morgan_orange> #info lab improvement => question raised to TSC - 50 K to upgrade lab capabilities 13:39:34 <morgan_orange> #info lab is already under constraints (no more Memory slot / free ports on switches,..) 13:40:06 <morgan_orange> #info ~ 3 servers + 1 switch should be compatible with the budget 13:40:32 <morgan_orange> #info other option would be to use another lab 13:41:12 <morgan_orange> #info Marco indicates that an upgrade of the lab could be a temp victory as, it is constantly growing, we risk to reach new resources frontiers in 2 or 3 releases 13:44:20 <morgan_orange> #info morgan_orange mentions that some resources had been purchased by LF for OPNFV - they had to support 5 Openstack installers in //, there are only 1 or 2 and muche less projects. It should be possible to recover resources if they are not used 13:50:47 <morgan_orange> #info Marco shares ecomp experience. system was smaller and simpler. 13:51:40 <morgan_orange> #info help deploy everything in the same namespace => could multi namespaces avoid that a component change breaks the whole cluster 13:52:46 <morgan_orange> #info other El Alto topic testsuite code improvement, CSIT code improvement, Use case Automation, Performance benchmark ... to be discussed offline 13:53:37 <morgan_orange> #topic Integration lab resource (put quota on tenants after we can provide shared infrastructure for project teams to test module deployment) 13:53:44 <morgan_orange> #info discussed in the previous section 13:54:04 <morgan_orange> #topic Benchmark planninBenchmark planning for EI Alto (China Telecom) 13:54:27 <morgan_orange> #info China Telecom performed performance tests on Beijing 13:55:10 <morgan_orange> #info Brian and integration team provided support 13:55:33 <morgan_orange> #info suggestion to move to Dublin 13:59:50 <morgan_orange> #info they use the vFW use case to assess performance 14:00:22 <morgan_orange> #action Lei Shi create a wiki page for performance tests (referenced from integration page) 14:00:43 <morgan_orange> #topic AoB 14:03:34 <morgan_orange> #info OOM stagging branch => no more manifest 14:04:48 <morgan_orange> #info tooling: https://gitlab.com/Orange-OpenSource/lfn/onap/python-onapsdk idea is to replace https://gitlab.com/Orange-OpenSource/lfn/onap/onap-tests 14:05:39 <morgan_orange> #endmeeting