========================================= #fdio-meeting: FD.io CSIT Project Meeting ========================================= Meeting started by mackonstan at 14:01:33 UTC. The full logs are available at http://ircbot.wl.linuxfoundation.org/meetings/fdio-meeting/2021/fd_io_csit_project_meeting/fdio-meeting-fd_io_csit_project_meeting.2021-09-29-14.01.log.html . Meeting summary --------------- * Vratko Polak. (vrpolak, 14:03:25) * Dave Wallace (dwallacelf, 14:06:33) * Fan Zhang (fan_zhang, 14:06:42) * Peter Mikus (pmikus, 14:07:17) * mackonstan (mackonstan, 14:07:31) * Tibor Frank (tifrank, 14:07:37) * Agenda bashing (mackonstan, 14:08:06) * Inputs from projects - VPP (mackonstan, 14:08:16) * rc1 branch pulled, on schedule (mackonstan, 14:08:24) * Inputs from projects - VSAP (mackonstan, 14:08:33) * Dave, no updates (mackonstan, 14:08:43) * Inputs from projects - TSC (mackonstan, 14:08:50) * Maciek, vendor funding agreement for fd.io lab got voted in, Ray raising a point about adding more TSC members, noting Arm (mackonstan, 14:11:15) * New tests, refactored tests (mackonstan, 14:11:37) * Fan, E810 flow "offload" vpp device tests, working with 4 rxq (didn't work with 8 rxq), unclear why it didn't work with 8 rxq. (mackonstan, 14:12:42) * Juraj Linkeš (jlinkes, 14:15:54) * Fan, looking to have both dpdk and avf drivers, for the same test, ~23 tests, adding ~5 min to job execution time (mackonstan, 14:16:14) * Juraj, re SPD flow cache patches, no update, found some inconsistencies of how the inbound and outbound rules are applied, confirmed with Neale, and now looking to submit a VPP patch to adjust/correct this (mackonstan, 14:18:01) * Viliam, TRex b2b loopback tests, all complted and merged now (mackonstan, 14:19:00) * Tibor, adding to weekly jobs execution and trending (mackonstan, 14:19:11) * Vratko, IP scale tests, spent considerable amount of time of trying to use VAT2, in vain, VAT2 is broken, so sticking to CLI. Notified Ole about VAT2 issues. (mackonstan, 14:21:13) * Vratko, note that IPsec 60k tunnel scale tests with VAT + exec are also failing due to the overall timeout … (mackonstan, 14:24:20) * Releases CSIT-2110 (mackonstan, 14:27:35) * Tibor, no major issues, few glitches, started rc1 tests, started dpdk and vpp 2106 retest after TG servers FVL fw update by Peter; first results expected Monday (mackonstan, 14:29:24) * Physical Infrastructure (mackonstan, 14:30:45) * Peter, finishing the upgrade of FVL firmwares and drivers on all TG servers, one NIC down tb24 awaiting Vexxhost onsite assistance to swap. (mackonstan, 14:32:26) * Peter, implemented 1GB hugepages, with selector of 2MB vs 1GB hugepages in vpp startup.conf (mackonstan, 14:33:31) * Peter, Maciek, Vratko, agreed to move ASAP/now ahead with FVL firmware and driver upgrades for SUTs for x86. (mackonstan, 14:39:20) * Juraj, trying it also on Arm testbeds (mackonstan, 14:40:14) * Maciek, missing ICX supermicro hw is not expected in the lab until mid November (mackonstan, 14:41:25) * Virtual Infrastructure, CI/CD (mackonstan, 14:41:36) * Dave, post mortem of last week outage completed, was due to routing misconfiguration and FD.io traffic using inter-site wireguard tunnel (mackonstan, 14:43:05) * Dave, no issues since then (mackonstan, 14:43:13) * Peter, Tibor, s3 content types and rendering fixed for read-the-docs (mackonstan, 14:44:03) * 33690: Docs: Discussion on oper branches | https://gerrit.fd.io/r/c/csit/+/33690 (vrpolak, 14:47:26) * Peter, Dave, gerrit links to s3 logs was fixed by Vanessa, but not entirely, url is still not resolving, Peter to follow up with Vanessa (mackonstan, 14:47:33) * Vratko, patch with proposal about "refreshed" operational branches is updated now based on comments, 33690: Docs: Discussion on oper branches | https://gerrit.fd.io/r/c/csit/+/33690 (mackonstan, 14:48:57) * LINK: https://gerrit.fd.io/r/c/csit/+/33690 (mackonstan, 14:49:04) * discussion about the merit of the proposal vs existing operational model (mackonstan, 14:57:15) * Environment, Framework (mackonstan, 14:58:11) * Trending (mackonstan, 14:58:20) * Maciek, few tests keep failing, mainly high scale nat44ed tests (mackonstan, 14:58:36) * Vratko, VPP AVF RSS fix is in place, results are visible in trending (mackonstan, 14:58:57) * Vratko, Maciek, should review small scale ipsec tests e.g. replace 4 tunnel tests with 40 tunnels … (mackonstan, 15:00:55) Meeting ended at 15:01:09 UTC. People present (lines said) --------------------------- * mackonstan (40) * collab-meetbot (4) * vrpolak (2) * dwallacelf (1) * fan_zhang (1) * pmikus (1) * tifrank (1) * jlinkes (1) Generated by `MeetBot`_ 0.1.4