#fdio-meeting: FD.io CSIT Project Meeting

Meeting started by mackonstan at 14:02:15 UTC (full logs).

Meeting summary

    1. Vratko Polak. (vrpolak, 14:02:49)

  2. Agenda bashing (mackonstan, 14:03:10)
  3. Inputs from projects - VPP (mackonstan, 14:03:22)
    1. Andrew - vpp 23.06 rc1 in progress (mackonstan, 14:03:51)
    2. Tibor - ready for rc1 dry run tests as soon as tomorrow (mackonstan, 14:04:22)

  4. Inputs from projects - TSC (mackonstan, 14:04:30)
  5. VPP and DPDK version (mackonstan, 14:05:12)
    1. Vratko - no changes related directly to dpdp ver change to 23.03 (mackonstan, 14:05:46)

  6. csit-2302 RCAs (mackonstan, 14:06:33)
    1. Vratko - vpp-2072 stuck with -1 from Damjan, suggestion to move it to vpp ethernet-input (mackonstan, 14:08:10)
    2. Vratko - vpp-2072 stuck with -1 from Damjan, suggestion to move it to "ethernet code (src/vnet/ethernet), not vnet_hw_interface." (mackonstan, 14:08:51)
    3. 38304: interface dpdk avf: introducing setting RSS hash key feature | https://gerrit.fd.io/r/c/vpp/+/38304 (mackonstan, 14:09:00)
    4. https://jira.fd.io/browse/VPP-2072 (mackonstan, 14:09:09)

  7. Next Release CSIT-2306 (mackonstan, 14:09:38)
    1. Vratko re reassembly tests - gtpu and ipsec already merged, gnv and vxlan require more work as it needs 3-node testbeds (mackonstan, 14:11:50)
    2. Vratko re mtu cleanup - applied to tests using dpdk drivers only. no jumbo frame tests for tests without dpdk drivers. (mackonstan, 14:13:38)
    3. Andrew reporting - issue with rc1 csit-vpp-verify-device for arm - queue buildup (mackonstan, 14:18:49)
    4. https://jenkins.fd.io/label/csit-builder-ubuntu2204-prod-aarch64/load-statistics (mackonstan, 14:18:52)
    5. Andrew considering to force merge it, just FYI, not expecting to impact rc1 artefacts (mackonstan, 14:19:31)
    6. Peter - 2n-spr, 3nb-spr - will be in 23.06 (mackonstan, 14:20:30)
    7. Peter - 3na-spr mlx cx7 - trex is working, vpp hitting issues - working with Benoit to t/shoot this, as neither cx6 nor cx7 where ever tested with vpp (mackonstan, 14:21:31)

  8. Physical Infrastructure (mackonstan, 14:22:04)
    1. Peter - looking for one missing cx6 card. cx6 placement into 3n-alt testbed can go ahead. (mackonstan, 14:23:08)
    2. Maciek re lab extension - SuperMicro SPR order will go ahead with Intel providing processors directly to SM before systems ship to Vexxhost faciliities (mackonstan, 14:26:40)
    3. Maciek re lab extension - ICX-D - finalizing updated specs with Intel (mackonstan, 14:27:01)

  9. Trending and Test Failures (mackonstan, 14:27:17)
    1. Vratko - observing systematic perf difference between testbeds 2n-icx, 3.2% difference for MRR tests (ndr pdr too) (mackonstan, 14:29:51)
    2. https://jira.fd.io/browse/CSIT-1915 - 2n-icx testbeds to not have the same performance (mackonstan, 14:30:03)
    3. Vratko - no latency reported by TRex when TRex using e810cq - due to missing functionality in ICE dpdk driver (mackonstan, 14:33:11)
    4. https://jira.fd.io/browse/CSIT-1914 - tests for some NICs do not produce latency data (mackonstan, 14:35:02)
    5. Kai - update on enabling DSA in SuperMicro servers - working with SM re new BIOS ver to enable it (mackonstan, 14:37:13)

  10. LFN Developer & Testing Forum 6-8 June (mackonstan, 14:38:40)
    1. csit.fd.io overview talk being scheduled for Thu 8June time likely 6-7am PDT / 3-4pm CEST (mackonstan, 14:39:45)


Meeting ended at 14:39:48 UTC (full logs).

Action items

  1. (none)


People present (lines said)

  1. mackonstan (35)
  2. collab-meetbot (9)
  3. vrpolak (1)
  4. dwallacelf (0)
  5. jlinkes (0)
  6. tifrank (0)


Generated by MeetBot 0.1.4.