#fdio-csit: FD.io CSIT meeting

Meeting started by mackonstan at 15:04:05 UTC (full logs).

Meeting summary

  1. Agenda bashing (mackonstan, 15:04:12)
    1. Tibor Frank (tifrank, 15:04:36)
    2. Vratko Polak. (vrpolak, 15:05:03)
    3. Jan Gelety (jgelety, 15:05:09)

  2. Physical labs, testbeds (mackonstan, 15:05:45)
    1. no issues, no other updates (mackonstan, 15:05:54)

  3. Project updates (mackonstan, 15:06:06)
    1. VPP - no updates (mackonstan, 15:06:12)
    2. Juraj Linkeš (jlinkes, 15:06:36)
    3. TSC - Vratko update: wikipedia page for VPP - content discussion (mackonstan, 15:06:37)
    4. CNCF CNF testbed - Peter - no updates - CNCF busy with Kubecon last week (mackonstan, 15:07:47)

  4. Releases (mackonstan, 15:08:07)
    1. no updates, will cover CSIT-2001 progress later in the call (mackonstan, 15:08:54)

  5. Operations (mackonstan, 15:09:16)
    1. Vratko - ci-mgmt verify jobs broken due to some new dependency (mackonstan, 15:09:49)
    2. Vratko - issue with 3n-tsh, compilation job issue, not enough memory on jenkins executor, fixed now (mackonstan, 15:10:44)

  6. VPP code performance (mackonstan, 15:11:05)
    1. Vratko - no new issues, made some progress with tsh issue, issue with vpp_Device tap tests failing bi-secting (mackonstan, 15:12:32)

  7. Developments (mackonstan, 15:16:26)
    1. Question from a person from Mellanox - "Olfar Avif(spelling?)": see email sent to one of the FD.io lists (mackonstan, 15:17:45)
    2. Questions related to UDPI - universal deep packet inspection: (mackonstan, 15:18:59)
    3. Q1: regexp engines in MLX SmartNICs. In the DPDK context. (mackonstan, 15:19:03)
    4. Q2: connection awareness separation from UDPI. (mackonstan, 15:19:36)
    5. (mackonstan, 15:25:41)
    6. https://lists.fd.io/g/udpi-dev/message/46 (mackonstan, 15:25:45)
    7. questions above were from:  Opher Reviv (mackonstan, 15:26:20)

  8. Framework - Migration to Python 3.x (mackonstan, 15:27:25)
    1. Jan updating: perf tests started to work (thx to Peter), the rest is being verified led by Peter. (mackonstan, 15:28:48)
    2. Jan updating: vpp_device issues, traffic_scripts are broken - t/shooting WIP ( taking over from Balaji) (mackonstan, 15:29:54)
    3. Jan updating: email updates re unfreezing CSIT master branch (mackonstan, 15:31:05)
    4. Nidhy: completed LispUtil.py and LispSetup.py libraries, is there anything else to convert? Jan: right now, not, but things may pop up. (mackonstan, 15:33:04)
    5. Tibor: PAL prepared, run and tested, no issues, but need to merge with all other changes from Jan and Peter (mackonstan, 15:35:55)
    6. Deprecated tests: DMM and TLDK are still to be removed, patches are prepared. (mackonstan, 15:38:22)

  9. TRex (mackonstan, 15:39:14)
    1. Vratko: sometimes results of trial measurement gives higher value, making CSIT scripts longer to finish. Behaviour is similar to the case when TRex is asked to send packets faster than NIC is supporting. Started to happen (mackonstan, 15:41:27)
    2. Supicion: started to happen after TRex got upgraded to v2.61 (mackonstan, 15:41:51)
    3. Unusual measurements coincide with TRex taking longer to complete a 1sec duration trial (mackonstan, 15:45:15)
    4. unusual measurments == 20% relative difference (mackonstan, 15:45:49)
    5. more frequent in 2n-dnv testbed (random in ca. 2% trials), 2n-skx testbed (srv6 tests) (mackonstan, 15:47:02)
    6. Vratko to ping trex mailer with headsup on the above issue (mackonstan, 15:48:59)


Meeting ended at 15:49:37 UTC (full logs).

Action items

  1. (none)


People present (lines said)

  1. mackonstan (39)
  2. collabot` (4)
  3. vrpolak (1)
  4. jgelety (1)
  5. jlinkes (1)
  6. tifrank (1)


Generated by MeetBot 0.1.4.