====================================== #fdio-csit: FD.io CSIT project meeting ====================================== Meeting started by mackonstan at 14:05:38 UTC. The full logs are available at http://ircbot.wl.linuxfoundation.org/meetings/fdio-csit/2020/fd_io_csit_project_meeting/fdio-csit-fd_io_csit_project_meeting.2020-09-23-14.05.log.html . Meeting summary --------------- * Dave Wallace (dwallacelf, 14:05:49) * Tibor Frank (tifrank, 14:05:52) * Agenda bashing (mackonstan, 14:07:01) * Jan Gelety (jgelety, 14:07:03) * Inputs from projects VPP (mackonstan, 14:07:23) * Dave, v20.09 rls moving forward (mackonstan, 14:07:33) * Andrew, "Cooldown and external tests" extended rc2 milestone today (mackonstan, 14:08:37) * Andrew will ping once the dpdk build patch is in gerrit 29057 is to commence csit rc2 tests (mackonstan, 14:17:11) * Inputs from projects VSAP (mackonstan, 14:17:35) * Inputs from projects TSC (mackonstan, 14:17:51) * Releases CSIT-2009 (mackonstan, 14:18:16) * Maciek, rc1 tests showed the same error as in trending, AVF interface initialisation failing from time to time (mackonstan, 14:19:37) * Peter, applied set of fixes on the CSIT side, trending is now clean; fixes ported to rls2009 (mackonstan, 14:20:36) * Peter, VPP from time to time is marking AVF interfaces operationally up even though there is no link (mackonstan, 14:21:28) * Tibor, rc2 tests now started, vpp builds with rc2~0 (mackonstan, 14:25:38) * Jan, adjusted weekly schedule: Saturday 00:00 UTC; 2n-clx, 2n-skx, 3n-hsw daily Mon, Wed, Fri; 3n-skx, 2n-dnv, 3n-dnv daily Mon, Thu; 3n-tsh daily Thu. (mackonstan, 14:30:34) * Jan, new tests nat44det are all done, final check before merging. (mackonstan, 14:31:34) * Vratko, new tests nat44ed cps tests, run into problems delaying completion of TRex calibration (polling TRex ASTF affects TRex behaviour). (mackonstan, 14:34:37) * Peter, new tests nat44ed stl uni-dir tests, merged, sharing the settings with nat44det. (mackonstan, 14:36:33) * Tibor, adding new graphs for nat44 tests into separate sections, as there are quite a few new tests. (mackonstan, 14:37:28) * Physical Infrastructure (mackonstan, 14:37:48) * New AMD testbed (mackonstan, 14:38:35) * Keesang, Peter has updated lab spec documentation, added ansible playbooks for AMD, and also created topology files (mackonstan, 14:39:25) * Peter, testbeds are now fully installed, calibration data documented, dry run tests are running fine confirming all physical install and cabling are good! (mackonstan, 14:40:50) * Peter, now need to create jenkins job specs for 2n-zen2 performance tests (mackonstan, 14:41:29) * Peter, created a clone for daily job_spec based on 2n-clx (mackonstan, 14:45:45) * Keesang, BIOS settings have been tuned, now need to see the first results. (mackonstan, 14:46:42) * Physical Infrastructure - Arm (mackonstan, 14:48:31) * Juraj, space and socket allocated, installation to be done (mackonstan, 14:48:42) * FD.io lab extension in 2021 (mackonstan, 14:51:24) * Maciek, see my email. Budgetary guesstimate provided to Trishan. (mackonstan, 14:52:20) * Virtual Infrastructure, CI/CD (mackonstan, 14:52:31) * currently Jenkins maintenance scheduled for 19-Oct (mackonstan, 14:52:57) * Environment, Framework (mackonstan, 14:53:20) * Performance Tests (mackonstan, 14:53:36) * New tests: async crypto, only with HW crypto QAT cards on 3n-hsw (mackonstan, 14:54:27) * TRex calibration with MLX (mackonstan, 14:55:14) * Vratko, after TRex upgrade to v2.82, the duration stretching got reduced to 1.3 sec @ 60Mpps per direction. Reducing it to 46 Mpps per direction most test trial do not have duration stretching, but there is some dependency on encapsulation (ip4, ip6, dot1q, …). (mackonstan, 14:59:49) * Vratko, looking to have a single rate for all encapsulation (mackonstan, 15:02:55) Meeting ended at 15:03:11 UTC. People present (lines said) --------------------------- * mackonstan (39) * collab-meetbot (4) * dwallacelf (1) * tifrank (1) * jgelety (1) Generated by `MeetBot`_ 0.1.4