#fdio-meeting: FD.io CSIT Project Meeting

Meeting started by mackonstan at 14:05:30 UTC (full logs).

Meeting summary

    1. Juraj LinkeŇ° (jlinkes, 14:05:35)
    2. Tibor Frank (tifrank, 14:05:36)
    3. mackonstan (mackonstan, 14:05:36)
    4. Vratko Polak. (vrpolak, 14:05:45)
    5. Fan Zhang (fan_zhang, 14:06:03)

  1. Agenda bashing (mackonstan, 14:06:23)
  2. Inputs from projects (mackonstan, 14:06:34)
    1. Andrew, all good, v22.06 shipped, now rc0 for v22.10 (mackonstan, 14:07:10)
    2. Andrew, will propose rls plan soon (mackonstan, 14:07:23)
    3. Vratko, TSC, discussion about new plugin for integrating Strongswan into VPP (mackonstan, 14:08:10)
    4. Fan, working on an integrated docker image, self-contained, for testing the Strongswan plugin in VPP (mackonstan, 14:08:55)

  3. E810 flow "offload" tests (mackonstan, 14:10:03)
    1. Tibor, added to job_specs, running for report (mackonstan, 14:10:13)

  4. VPP and CSIT hot items (mackonstan, 14:11:03)
    1. Maciek, will re-visit both items after the rls2206 (mackonstan, 14:11:25)
    2. Maciek, will reach out to AMD team pro-actively, using rls2206 results as a reference (mackonstan, 14:12:29)

  5. CSIT-2206 (mackonstan, 14:12:42)
    1. Tibor, Mr Jenkins re-started work on Friday evening UTC, after recovering from shutdown mode (mackonstan, 14:13:29)
    2. Tibor, quite a few iterative runs completed, 4..10, depending on testbed (zn2 altra .. dnv icx skx) (mackonstan, 14:14:35)
    3. Tibor, current draft report in master showing first rls test data (mackonstan, 14:15:30)
    4. https://s3-docs.fd.io/csit/master/report/ (mackonstan, 14:15:34)
    5. Tibor, if CI/CD infra stays up, all iterative tests, and almost all coverage data will be available in time for rls date on Wed 13-Jul (mackonstan, 14:16:52)
    6. Tibor, new report version generated once a day (mackonstan, 14:17:17)
    7. Tibor, added frequency normalized comparison as last column to tables: 2n-icx vs 2n-clx, 3n-icx vs 3n-skx, 2n-icx vs 2n-skx, 2n-icx vs 2n-skx (mackonstan, 14:19:22)
    8. Maciek, Andrew, Fan, does it make sense to add comparison between different CPU architectures, Xeon flavours vs Atom flavours, vs Arm flavours - all agree this does not make sense (mackonstan, 14:21:59)
    9. Tibor, static content is ready, 3n-alt diagram to be added (Maciek, Juraj), rls notes, known issues and regressions to be updated (mackonstan, 14:24:16)
    10. Vratko, did one pass through latest rls data, so far aligned with rc2 (mackonstan, 14:25:30)
    11. Vratko, some issues with ASTF results (mackonstan, 14:25:45)
    12. Vratko, performance progressions for dpdk tests, as dpdk improvement was not cherry-picked in v22.02 (mackonstan, 14:29:49)
    13. Fan, dpdk v22.07 releasing in one week (mackonstan, 14:30:54)
    14. Vratko, re known issues, TRex ASTF TCP 1518 (mackonstan, 14:34:05)
    15. Vratko, need to add to known issues in the patch tracking the rls notes (mackonstan, 14:36:19)
    16. 36499: docs(report): Release notes 22.06 updated | https://gerrit.fd.io/r/c/csit/+/36499 (mackonstan, 14:36:21)
    17. Vratko, Maciek, some issues where known in previous release, so should be moved from New to Previous (mackonstan, 14:37:56)
    18. Fan, Maciek, Tibor, need to look at the e810 flow tests, Tibor to provide links to the report (mackonstan, 14:39:38)
    19. Tibor, will provide links to csit.fd.io UTI dashboard (mackonstan, 14:40:03)
    20. http://csit.fd.io/report/ (mackonstan, 14:42:22)
    21. http://csit.fd.io/report/#eNrtkcEOgjAMhp-GkymBKeDFg-h7GNwqLCo220D06R3GWIx68MLJy5bs-7u_7W_dyeDG4mERJHmQ5YHItPJHMF1O_CWtdkJEKbREIEQYpaE5WJjWoGUHIilRUIzzOOq6FhSpPWiabVxT1-hl6WwLsQR0lX8tHTXVuefbwiIcjeltxKq38X--ezJWjWPsO3khLRqGzxZZQtWF-ffGuaAwWHDFYB6WOLSDjj7Nx9qdKY5o9RW5wO-FufQBMIrlq4u70IA-lpat74oR06rVP62f03osbfy06J_W72nRM61kfQOvdmnU (mackonstan, 14:42:31)
    22. https://s3-logs.fd.io/vex-yul-rot-jenkins-1/csit-vpp-perf-report-iterative-2206-3n-icx/49/report.html.gz (mackonstan, 14:42:56)
    23. Vratko, Maciek multi-core speedup graphs for 100 GbE look good and as expected, except 2n-zn2 (new ticket), l2xc faster than testpmd on 4c on some testbeds (mackonstan, 14:48:46)

  6. Physical Infrastructure (mackonstan, 14:53:13)
    1. Maciek, Peter, need to think about re-purposing / decommissioning the 2n-skx and 3n-skx testbeds (mackonstan, 14:54:17)

  7. Virtual Infrastructure, CI/CD (mackonstan, 14:54:47)
    1. Maciek, Tibor, it is still unclear why LF IT FD.io Jenkins was put into shutdown mode (mackonstan, 14:55:39)
    2. Maciek, expecting to have this visited in TSC call tomorrow (mackonstan, 14:56:01)

  8. Environment, Framework (mackonstan, 14:56:23)
  9. Trending (mackonstan, 14:56:35)
    1. https://jira.linuxfoundation.org/plugins/servlet/theme/portal/2/IT-24238 <-- LFN ticket re Jenkins in shutdown mode. Not much about causes, more about what can we do to prevent it in case it has a non-human cause. (vrpolak, 14:57:46)
    2. Maciek, nobody reporting any new issues, will check with Viliam re CSIT failing perf tests for this week (mackonstan, 14:57:47)


Meeting ended at 15:00:23 UTC (full logs).

Action items

  1. (none)


People present (lines said)

  1. mackonstan (46)
  2. collab-meetbot` (3)
  3. vrpolak (2)
  4. jlinkes (1)
  5. tifrank (1)
  6. fan_zhang (1)


Generated by MeetBot 0.1.4.