15:01:08 #startmeeting 15-Nov FD.io CSIT Project Meeting 15:01:08 Meeting started Wed Nov 15 15:01:08 2017 UTC. The chair is mackonstan. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:01:08 Useful Commands: #action #agreed #help #info #idea #link #topic. 15:01:08 The meeting name has been set to '15_nov_fd_io_csit_project_meeting' 15:01:13 #chair 15:01:13 Current chairs: mackonstan 15:01:17 #info Dave Wallace 15:01:18 #info Tom Herbert 15:01:31 #info Tibor Frank 15:01:40 #info Nicolas Bouthors 15:02:00 #info Peter Mikus 15:02:24 #info Jan Gelety+ 15:02:38 #undo 15:02:45 #info Jan Gelety 15:03:35 #topic Agenda bashing 15:03:47 #topic Inputs from LF and FD.io projects 15:04:47 #info VPP - agreed on next rls to be 18.01 15:04:55 #link https://wiki.fd.io/view/Projects/vpp/Release_Plans/Release_Plan_18.01 15:07:08 #info HoneyComb - no new tests.. 15:08:17 #info need to check the plan for HC tests in CSIT with HC project lead - Marek Gradzki 15:09:19 #info TSC - new project approved - pma_tools 15:09:39 #link https://wiki.fd.io/view/Project_Proposals/pma_tools 15:11:22 #topic Operational status 15:15:23 #info jgelety - Centos7 semi-weekly job failed - 2 iterations worked, 3rd failed - identified pattern: vpp ip probe, nested qemu startup failures, can't get vpp pid at startup - summary with failed jobs into csit-dev 15:15:37 #info Tom offering to compare the failures with what's observed in his labs 15:17:45 #info pmikus - trending jobs - mthread vhost tests started to fail as 11/11 15:18:16 #topic rls1710 release 15:18:30 #link https://jenkins.fd.io/view/csit/job/csit-vpp-perf-trend-daily-master/191/ VHOST multithread failures 15:19:13 #info report reissued with detailed results 15:19:17 #link https://docs.fd.io/csit/rls1710/report/ 15:19:31 #topic next release - rls1801? 15:19:58 #info initial plan wiki page set up 15:20:00 #link https://wiki.fd.io/view/CSIT/csit1801_plan 15:23:02 #topic Presentation and Analytics Layer (PAL); 15:23:58 #info PAL additions for next releases: telemetry data inputs - vpp show run, results comparisons, multi-core scale-up analysis 1c/2c/4c 15:24:34 #topic Performance tests 15:28:04 #info jgelety - SW cryptodev perf tests are in hand 15:29:43 #info tifrank - SRv6 - high level design done and reviewed - ready for csit coding.. 15:30:54 #info dwallacelf - vpp tcp/ip hoststack should be ready for perf testing 15:31:26 #info pmikus - agree that tcp/ip hoststack is a reasonable target for experimental onboarding onto csit 15:32:10 #topic VPP functional tests - virl 15:32:18 #info jgelety - SW cryptodev requires GCM cipher support - requires scapy and pycrypto library upgrades. 15:32:23 #info jgelety - Ubuntu works fine, new nested-VM Centos7 image doesn't work (old one works fine). 15:33:03 #topic Test environment updates 15:33:15 #info lab extensions with XEON Skylake servers 15:35:50 #info mackonstan proposal: proposal to add 4 performance testbeds: 2x 2-node testbeds (for no encaps), 2x 3-node testbeds (for with encaps) 15:42:04 #info mackonstan - will send email with proposed high-level specs 15:42:18 #topic adding test equipment: Netgate 15:43:10 #info tfherbert - access to devices is finally sorted 15:45:13 #info total of 3 netgate devices available 15:45:59 #info need to decide which tests and what topologies 15:46:16 #info tfherbert to send initial thoughts to csit-dev 15:46:39 #topic ARM additions to CSIT labs 15:48:17 https://wiki.fd.io/view/VPP/AArch64#Meeting_Details 15:49:09 #info kumar reporting - ARM VPP issues tracked in jira 15:49:15 #link https://wiki.fd.io/view/VPP/AArch64 15:50:22 #info ThunderX servers shipped - ETA tbc 15:51:29 #topic NestedVM refresh - done? 15:51:48 #info ubuntu working, Centos7 - see tfherbert email report on csit-dev 15:52:48 #topic VIRL platform infra 15:57:13 #info tb4-virl3 - looking healthy, used for cryptodev SW test onboarding, once done, back into produciton 16:00:22 #info dwallacelf - re show run api - some code there, but not used, pointers into csit-dev mailer. 18:24:22 anyone know whom is behind the GENEVE test cases? they never pass on container builds and id love to figure out why that is happening 20:05:14 you mean in make test, y ? 20:22:35 sure..although that is being run as part of make verify 08:37:33 git blame 20:44:07 #endmeeting