#opnfv-vswitchperf: OPNFV VSPERF

Meeting started by ctrautma at 15:03:27 UTC (full logs).

Meeting summary

    1. Al Morton (acm, 15:05:27)

  2. T-Rex Integration (ctrautma, 15:05:59)
  3. T-Rex Integration (ctrautma, 15:06:13)
    1. Martin Goldammer (ctrautma, 15:07:02)
    2. Martin giving demo (ctrautma, 15:07:22)
    3. enabling Latency collection lowers final results (ctrautma, 15:16:49)
    4. known issue about T-Rex, work around propossed (ctrautma, 15:24:41)
    5. issue listed in T-Rex FAQ (ctrautma, 15:25:08)
    6. numbers between Ixia and T-Rex on same system are very similar (ctrautma, 15:29:51)
    7. current patch includes tput code for rfc 2544 (ctrautma, 15:30:23)
    8. patch needs final review from community (ctrautma, 15:33:31)
    9. Christian to do final review by end of the week (ctrautma, 15:34:20)

  4. test-api patch (ctrautma, 15:34:29)
    1. patch merged (ctrautma, 15:35:15)
    2. testcase names updated (ctrautma, 15:35:20)
    3. waiting on Jenkins CI job execution completion (ctrautma, 15:35:43)

  5. CI-Pod Tuning (ctrautma, 15:42:22)
    1. found issue with tuning core/ht when using 0 (ctrautma, 15:44:37)
    2. christian to check jira ticket (ctrautma, 15:46:13)
    3. results are 7 Mpps consistently with core/ht pairing (ctrautma, 15:46:44)
    4. Christian to run final longer duration and then have Martin try on active pod (ctrautma, 15:47:02)

  6. Traffic Generator Testing (ctrautma, 15:47:21)
    1. Al Morton (ctrautma, 15:47:46)
    2. compare results between Ixia Prox and Moongen (ctrautma, 15:48:24)
    3. Ixia changes destination L2 (ctrautma, 15:49:18)
    4. Prox and Moongen change L3 (ctrautma, 15:49:27)
    5. Prox changes source and destination L3 (ctrautma, 15:49:41)
    6. Moongen changes source L3 (ctrautma, 15:49:54)
    7. Martin also notices differences when doing Multistream at different layers (ctrautma, 15:52:49)
    8. VPP not as sensitive to different multistream profiles (ctrautma, 15:53:05)
    9. Trevor inquire Ixia should send patch about more multistream capabilities (ctrautma, 15:55:10)
    10. Is higher flow counts realistic for real use case profiles? (ctrautma, 15:56:56)

  7. Documentation (ctrautma, 16:00:03)
    1. Martin Klozik (ctrautma, 16:00:09)
    2. community needs to review patches (ctrautma, 16:00:23)
    3. all involved parties and need to review their pieces for E release (ctrautma, 16:00:56)


Meeting ended at 16:01:22 UTC (full logs).

Action items

  1. (none)


People present (lines said)

  1. ctrautma (37)
  2. collabot (4)
  3. acm (1)


Generated by MeetBot 0.1.4.