#fdio-meeting: FD.io CSIT Project Meeting

Meeting started by mackonstan at 13:58:27 UTC (full logs).

Meeting summary

    1. Tibor Frank (tifrank, 14:02:09)

  1. Agenda bashing (mackonstan, 14:04:17)
    1. Juraj Linkeš (jlinkes, 14:04:24)

  2. Inputs from projects - VPP (mackonstan, 14:04:31)
    1. mackonstan (mackonstan, 14:04:35)
    2. Andrew: rc2 in progress, artefacts ready in an hour or so (mackonstan, 14:04:58)

  3. Inputs from projects - TSC (mackonstan, 14:05:45)
    1. LFN Casey - asking for FD.io CSIT lab inventory, Maciek took AI to open vexxhost ticket enquiring if Vexxhost has inventory record with S/Ns (mackonstan, 14:12:05)

  4. New or refactored tests - Scale tests API optimization (mackonstan, 14:12:41)
    1. Vratko - got job execution time data for 2n-clx, it's >12hrs, working out which high-scale test to remove to fit the time budget (mackonstan, 14:14:17)

  5. move to Ubuntu 22.04 (mackonstan, 14:15:03)
    1. Peter, vpp_device so far staying on Ubuntu 20.04 (mackonstan, 14:15:36)
    2. Peter, 3n-snr - trying to get it working with TRex + CVL, packet loss got solved with TRex upgrade v3.00 (mackonstan, 14:17:28)
    3. Peter, 3n-snr - unfortunately 25GbE interface are randomly going down, reached out to Vexxhost on-site support to re-seat cables. Didn't help. (mackonstan, 14:18:50)
    4. Peter, to make progress need more direct Intel assistance. work with Maciek on this (mackonstan, 14:20:24)

  6. Next Release CSIT-2210 (mackonstan, 14:21:00)
    1. Peter, re TRex version upgrade to address CVL (partial) compatibility and 3n-snr, not enough data to make a call if TRex v3.00 is good enough (mackonstan, 14:25:12)
    2. Vratko, question is if we have enough time to move to TRex v3.00, and if things go bad and south, revert (mackonstan, 14:25:45)
    3. Maciek, agree with that (mackonstan, 14:25:50)
    4. Tibor, we should have enough time to do it, including the worst case scenario of reverting TRex sw version (mackonstan, 14:26:35)
    5. Peter, upgrading TRex - following steps: i) automated upgrade of TRex - quick; ii) manual firmware upgrades (not automated for various reasons) - takes time; iii) can do 2n-clx, run rc2/trending tests, then evaluate quickly; iv) proceed with other testbed types in parallel to step (iii) (mackonstan, 14:35:27)
    6. Maciek, agreed above between Tibor, Peter, Vratko and Maciek (mackonstan, 14:36:07)
    7. Juraj, Arm testbeds still down, patch is still in review queue, no feedback (mackonstan, 14:37:33)
    8. Andrew, should send email if no replies in gerrit (mackonstan, 14:37:58)
    9. https://gerrit.fd.io/r/c/vpp/+/37199 (jlinkes, 14:39:04)
    10. Maciek, this patch is on a critical path for Arm tests to ship in csit-2210 (mackonstan, 14:39:18)
    11. Tibor, Maciek - we have 2 weeks for all rc2 testing issues per above (mackonstan, 14:40:25)

  7. Physical Infrastructure (mackonstan, 14:40:46)
    1. Maciek, waiting for Nvidia/MLX contact to confirm the qty and ETA for the new MLX 2p100GbE NICs (cx6) (mackonstan, 14:42:09)

  8. Virtual Infrastructure, CI/CD (mackonstan, 14:42:48)
    1. Vratko, re VPP API changes' tracking - check info is available in api crc check job, but need to start making use of it, specifically provide a feedback loop to vpp-dev (there should be no in-progress or deprecated api's in CSIT) (mackonstan, 14:50:24)

  9. Trending (mackonstan, 14:56:10)
    1. Viliam reporting on failures, see updated H/M/L priority labels posted on meeting wiki page (mackonstan, 14:56:48)


Meeting ended at 15:01:04 UTC (full logs).

Action items

  1. (none)


People present (lines said)

  1. mackonstan (33)
  2. collab-meetbot (8)
  3. jlinkes (2)
  4. tifrank (1)
  5. vrpolak (0)


Generated by MeetBot 0.1.4.