========================================= #fdio-meeting: FD.io CSIT Project Meeting ========================================= Meeting started by mackonstan at 14:04:07 UTC. The full logs are available at http://ircbot.wl.linuxfoundation.org/meetings/fdio-meeting/2023/fd_io_csit_project_meeting/fdio-meeting-fd_io_csit_project_meeting.2023-08-09-14.04.log.html . Meeting summary --------------- * Tibor Frank (tifrank, 14:04:18) * Maciek Konstantynowicz (mackonstan, 14:04:27) * Agenda bashing (mackonstan, 14:04:48) * Inputs from projects (mackonstan, 14:05:10) * Dave v23.10 dates approved on vpp-dev call yday (mackonstan, 14:05:53) * LINK: https://gerrit.fd.io/r/c/vpp/+/39299 (vrpolak, 14:07:26) * Juraj Linkeš (jlinkes, 14:07:51) * Dave vpp-dev call yday - discussed Vratko's issue re IPsec async scheduler (crypto nodes invoked on vpp main worker): see Vratko's fix above to vpp code (mackonstan, 14:07:58) * Dave Wallace (dwallacelf, 14:08:23) * LINK: 39299: crypto-sw-scheduler: avoid crypto work on vpp_main | https://gerrit.fd.io/r/c/vpp/+/39299 (mackonstan, 14:08:42) * above fixes issue tracked in trending issue register (mackonstan, 14:10:27) * LINK: https://wiki.fd.io/view/CSIT/TestFailuresTracking#.28H.29_ipsec_swasync:_VPP_can_become_unresponsive (mackonstan, 14:10:29) * tracked in jira (mackonstan, 14:10:35) * LINK: https://jira.fd.io/browse/CSIT-1928 (mackonstan, 14:10:48) * Dave - discussed timing for the meeting, Kai suggested to keep for now until next meeting, as more Intel team may be coming online in East Asia (mackonstan, 14:13:19) * Inputs from projects - TSC (mackonstan, 14:13:48) * Dave - onboarding a new LFN intern (mackonstan, 14:14:00) * New or refactored tests (mackonstan, 14:16:38) * VPP and CSIT (not so hot) items (mackonstan, 14:17:02) * re HW accelerators - Peter - have live debug sessions with Kai and the Intel dev team - making progress (mackonstan, 14:17:47) * Previous Release CSIT-2306 (mackonstan, 14:18:23) * 2306 RCAs - Vratko - no big updates. will cover on jira ticket updates once we get to trending issue (mackonstan, 14:20:33) * new release CSIT-2310 (mackonstan, 14:21:21) * schedule now confirmed, per vpp-dev, csit-dev is aligned, dates in wiki link below (mackonstan, 14:21:39) * LINK: https://wiki.fd.io/view/CSIT/csit2310_plan (mackonstan, 14:21:41) * re Patch for VPP moving to dpdk 23.07 - see notes from last week meeting (mackonstan, 14:24:23) * "Kai - agree. Already working on bumping dpdk version patch to dpdk v23.07. ETA mid August. (mackonstan, 14:18:35)" (mackonstan, 14:24:46) * LINK: https://ircbot.wl.linuxfoundation.org/meetings/fdio-meeting/2023/fd_io_csit_project_meeting/fdio-meeting-fd_io_csit_project_meeting.2023-08-02-14.07.html (mackonstan, 14:25:05) * LINK: https://gerrit.fd.io/r/c/vpp/+/39352 (dwallacelf, 14:25:34) * LINK: 39352: dpdk: bump to DPDK 23.07 | https://gerrit.fd.io/r/c/vpp/+/39352 (mackonstan, 14:26:13) * vpp and csit 2310 rc1 is 2023-09-20 => 6 weeks ago (mackonstan, 14:27:06) * Physical Infrastructure (mackonstan, 14:28:53) * Dave - no status on new orders or to be approved orders (mackonstan, 14:29:04) * Peter - there is update re new order - will ping Dave for sync (mackonstan, 14:31:01) * Maciek - SuperMicro should be putting SPR processors received from Intel before shipping the systems to Vexxhost (mackonstan, 14:31:25) * 4 SPRs received by Vexxhost yesterday - see email from Ali @Vexxhost "This email is to inform you that we received 4x supermicro servers at DC" (mackonstan, 14:33:03) * Virtual Infrastructure, CI/CD (mackonstan, 14:33:26) * Environment, Framework (mackonstan, 14:33:50) * Vratko - for in-progress msgs, Andrew created patch per vpp component, waiting for few weeks to collect comments, and then merge them (mackonstan, 14:34:30) * Vratko to pasted some sample urls for gerrit patches (mackonstan, 14:34:48) * Trending and Test Failures (mackonstan, 14:35:01) * Vratko - sw async already covered earlier (mackonstan, 14:35:24) * Vratko - wireguard issues (mackonstan, 14:36:19) * LINK: https://wiki.fd.io/view/CSIT/TestFailuresTracking#.28M.29_3n-icx.2C_3n-snr:_wireguard_100_and_1000_tunnels_mlrsearch_tests_failing_with_2c_and_4c (mackonstan, 14:36:36) * LINK: https://jira.fd.io/browse/CSIT-1886 | 3n-icx, 3n-snr: Wireguard tests with 100 and more tunnels are failing PDR criteria (mackonstan, 14:36:56) * Vratko, Maciek - discussion - wireguard rekeying on vpp introduces packet loss (mackonstan, 14:42:04) * LINK: https://git.fd.io/vpp/tree/MAINTAINERS#n741 (mackonstan, 14:42:54) * agreed to check with the maintainers if this is per design and current code, and if so, how CSIT should test it (mackonstan, 14:43:21) * if it is a bug, path to have it fixed (mackonstan, 14:43:32) * Vratko - 2ndary issue - no useful error counters for wireguard showing this problem (mackonstan, 14:45:26) * Vratko - making it very very hard to investigate (mackonstan, 14:45:40) * Analytics, Presentation (mackonstan, 14:46:32) * Tibor - upgrade of py libraries powering csit.fd.io is almost done, dev complete, in testing (mackonstan, 14:47:34) * Peter - some "wrinkles" still to solve after the upgrade. more in next meeting. (mackonstan, 14:48:22) Meeting ended at 14:48:55 UTC. People present (lines said) --------------------------- * mackonstan (53) * collab-meetbot` (9) * dwallacelf (2) * tifrank (1) * jlinkes (1) * vrpolak (1) Generated by `MeetBot`_ 0.1.4