#fdio-meeting: FD.io CSIT Project Meeting

Meeting started by mackonstan at 15:01:20 UTC (full logs).

Meeting summary

    1. mackonstan (mackonstan, 15:01:27)
    2. Tibor Frank (tifrank, 15:01:31)
    3. Vratko Polak. (vrpolak, 15:01:45)

  2. Agenda bashing. (mackonstan, 15:02:53)
  3. Inputs from projects - VPP (mackonstan, 15:03:22)
    1. Andrew - rc2 artefacts in packagecloud, no known open issues (mackonstan, 15:03:59)

  4. Inputs from projects - VSAP (mackonstan, 15:04:27)
    1. Dave Wallace (dwallacelf, 15:04:28)
    2. Dave, haven\t heard from the project for a long time … will check with Florin (mackonstan, 15:04:43)

  5. TSC updates (mackonstan, 15:05:12)
    1. Dave, discussed MLX NICs, Ed is pinging MLX re donations (mackonstan, 15:05:53)
    2. Dave, opened vexxhost ticket to add RAM to ampere build servers (mackonstan, 15:06:52)

  6. E810 flow "offload" tests (mackonstan, 15:09:25)
    1. Fan, perf test are being worked on (mackonstan, 15:09:37)
    2. Fan, device tests to be done later (mackonstan, 15:09:47)

  7. Scale tests (mackonstan, 15:10:13)
    1. Vratko, some things in works, but not focusing on it right now (mackonstan, 15:10:25)
    2. Tibor, Address perf test jobs with excessive execution time - weekly ndrpdr, just emailed the latest analysis, 2n-skx/clx/icx; icx the slowest, under investigation as to what the RCA is. (mackonstan, 15:12:32)

  8. VPP RC1, RC2 hot patch(es) items (mackonstan, 15:13:18)
    1. MTU, Vratko, several issues: i) CSIT setting high value (9200B) in hw_interface_set_mtu call, resulting in config error; ii) CSIT setting values in hw_interface_set_mtu for VPP interfaces that don't support hw mtu. Reason behind approach i) was that in the past DPDK was performing worse when setting MTU < 9200B. (mackonstan, 15:17:34)
    2. re MTU, Vratko, Maciek, agreed to remove hw_interface_set_mtu call from the csit tests, and instead rely on VPP default settings and behaviour. (mackonstan, 15:26:16)
    3. Maciek, note that "MTU" in above comments in fact means "Maximum Frame Size" as specified in IEEE 802.3-2008 (mackonstan, 15:37:21)
    4. Maciek, Vratko, need to add a new agenda item, tracking "Maximum Frame Size" handling by drivers, VPP (and CSIT if needed) (mackonstan, 15:44:55)
    5. Peter, note that after DPDK bump and related VPP changes 2n-dnv and 3n-dnv are broken due to 1GE auto-negotiating on 10GE niantic . Fan is happy to help. (mackonstan, 15:48:16)
    6. Peter, is niantic supported in dpdk, Fan to follow up. (mackonstan, 15:50:28)

  9. Releases CSIT-2202 (mackonstan, 15:50:46)
    1. Tibor, rc1 is done. Test data in master ver of report. incl. 2n-icx and 3n-icx. (mackonstan, 15:51:26)
    2. Tibor, rc2 testing starting tomorrow. (mackonstan, 15:51:52)
    3. Tibor, no unexpected test failures. (mackonstan, 15:52:53)

  10. Physical Infrastructure (mackonstan, 15:53:25)
    1. no issues (mackonstan, 15:54:18)

  11. Virtual Infrastructure, CI/CD (mackonstan, 15:54:48)
    1. Dave, nothing new (mackonstan, 15:55:02)
    2. Dave, arm vpp-device jobs are currently not run per vpp patch (mackonstan, 15:56:31)
    3. Peter, will run csit-vpp-device tests on arm for csit-2202 rls (mackonstan, 15:59:40)

  12. Trending (mackonstan, 16:00:10)
    1. Tibor, added 2n-icx 3n-icx testbeds (mackonstan, 16:00:28)
    2. Vilaim, no new failures in trending (mackonstan, 16:00:45)


Meeting ended at 16:00:48 UTC (full logs).

Action items

  1. (none)


People present (lines said)

  1. mackonstan (37)
  2. collab-meetbot` (3)
  3. tifrank (1)
  4. vrpolak (1)
  5. dwallacelf (1)


Generated by MeetBot 0.1.4.