#fdio-meeting: FD.io CSIT Project Meeting

Meeting started by mackonstan at 14:00:20 UTC (full logs).

Meeting summary

    1. Tibor Frank (tifrank, 14:02:41)

  2. Agenda bashing (mackonstan, 14:03:12)
    1. Juraj Linkeš (jlinkes, 14:03:25)
    2. viliam luc (viliamluc, 14:03:36)
    3. proposal: Dave Wallace has a standing collision with other FD.io meetings, proposal to move it to 6:30-7:30am pacific. (mackonstan, 14:03:55)
    4. MK topic proposal: are 2n- and 3n-dnv results reliable and report worthy, need to decide if keeping them for csit-2210 report while still running on ubuntu 20.04 (mackonstan, 14:05:16)

  3. Inputs from projects - VPP (mackonstan, 14:06:24)
    1. Vratko Polak. (vrpolak, 14:07:19)
    2. Andrew: Dave has launched a make test survey - current usage of various make test features, see pointer in vpp-dev mailer (mackonstan, 14:07:28)

  4. Inputs from projects - TSC (mackonstan, 14:07:41)
  5. New or refactored tests (mackonstan, 14:08:55)
  6. Scale tests API optimization (mackonstan, 14:09:40)
    1. Vratko: still need to run the test & jobs to compare durations as discussed on the last week call (mackonstan, 14:10:40)

  7. CSIT env move to Ubuntu 22.04 (mackonstan, 14:11:40)
    1. status: all done except perf dnv and vpp_device (mackonstan, 14:13:00)

  8. Lab new builds: 3n-snr (mackonstan, 14:14:30)
    1. Peter, Maciek: 3n-snr now operational, but producing a mixed-bag of results, due to using unsupported TRex + CVL combination (mackonstan, 14:15:58)
    2. Peter, Maciek: proposal to upgrade TRex to v2.99 that offers initial (incomplete) support for CVL ICE DPDK driver. ETA unknown. (mackonstan, 14:17:00)

  9. rls csit-2210 (mackonstan, 14:18:45)
    1. Tibor, re-run soak tests, working now, after fix(es) from Vratko (mackonstan, 14:19:45)
    2. Tibor, rc1 tests, hoststack run on 3n-icx, all vpp echo fail (udp, quic), all other tests pass (mackonstan, 14:21:20)
    3. Viliam, the same hoststack vpp echo tests were also failing in csit-2206, see ticket in known issues (mackonstan, 14:22:05)
    4. VPPecho failing. Ticket: https://jira.fd.io/browse/CSIT-1835 (viliamluc, 14:22:32)
    5. Tibor, we won't do the re-run of vpp v22.06 in csit-2210 env, as no ubuntu 22.04 images available for vpp v22.06 (mackonstan, 14:24:53)
    6. https://gerrit.fd.io/r/c/vpp/+/37199 (jlinkes, 14:25:12)
    7. 37199: vppinfra: fix set mempolicy for different cpu arch | https://gerrit.fd.io/r/c/vpp/+/37199 (mackonstan, 14:27:01)
    8. Juraj: patch needs to be reviewed and merged (mackonstan, 14:28:13)
    9. Tibor, Maciek: re csit-2210 report - major change is that there won't be any static html graphs with results data anymore, instead we will have url links to csit.fd.io perf dashboard (mackonstan, 14:31:11)
    10. Tibor: static content refactor by Peter for methodology section - available in draft master report (mackonstan, 14:32:48)
    11. Tibor: re csit-2210 report - correction, there will be static html graphs in the report. (mackonstan, 14:35:19)
    12. Tibor: will compile a report TODO list in the next week as per usual (mackonstan, 14:35:56)
    13. Tibor: adding 3n-snr (mackonstan, 14:36:09)

  10. Physical Infrastructure (mackonstan, 14:36:47)
  11. Virtual Infrastructure, CI/CD (mackonstan, 14:37:05)
  12. Environment, Framework (mackonstan, 14:37:13)
    1. Maciek: only change is TRex v2.99 testing with CVL in 3n-snr testbed (mackonstan, 14:37:33)

  13. Trending (mackonstan, 14:37:54)
    1. Viliam - no updates from last week (mackonstan, 14:39:42)
    2. Viliam, Vratko: re issue "4. 2n-clx: half of the packets lost on PDR tests (re-opened). CSIT-1864" - detected in weekly run, doesn't happen always, looks like HW dependency, unclear if it's always on the same testbed (mackonstan, 14:41:52)
    3. Viliam: confirmed that CSIT-1864 failures were observed on 3 different 2n-clx testbeds: .19 .23 29 (mackonstan, 14:44:26)
    4. Viliam: confirmed that CSIT-1864 failures were observed on 3 different 2n-clx testbeds: .19 .21 29 (CORRECTION) (mackonstan, 14:44:57)

  14. CSIT meeting time change proposal (mackonstan, 14:47:19)
    1. proposal: Dave Wallace has a standing collision with other FD.io meetings, proposal to move it to 6:30-7:30am pacific. (mackonstan, 14:47:22)
    2. Juraj: have another meeting 0.5hr earlier, so have collision with 6:30-7:30am pacific. will check if another meeting can be moved and report back. (mackonstan, 14:50:17)

  15. 2n-dnv and 3n-dnv results reliability (mackonstan, 14:50:51)
    1. Viliam, Maciek: are 2n- and 3n-dnv results reliable and report worthy, need to decide if keeping them for csit-2210 report while still running on ubuntu 20.04 (mackonstan, 14:51:04)
    2. Maciek: let's get the links for csit.fd.io 2n-dnv and 3n-dnv, and decide based on this, can do over email on csit-dev (mackonstan, 14:51:51)

  16. Other items (mackonstan, 14:53:53)
    1. Kai: asking if there is anything to be tracked from Intel side (mackonstan, 14:58:20)
    2. Maciek: Yulong Pei reported 3 patch sets with new tests that he would like to have merged for csit-2210 release - see email thread on csit-dev. one of them is merged last time i checked. (mackonstan, 14:58:56)
    3. https://lists.fd.io/g/csit-dev/message/4541 (mackonstan, 14:59:33)


Meeting ended at 14:59:36 UTC (full logs).

Action items

  1. (none)


People present (lines said)

  1. mackonstan (48)
  2. collab-meetbot (7)
  3. jlinkes (2)
  4. viliamluc (2)
  5. vrpolak (1)
  6. tifrank (1)


Generated by MeetBot 0.1.4.