#fdio-csit: FD.io CSIT project meeting

Meeting started by mackonstan at 14:02:23 UTC (full logs).

Meeting summary

    1. Vratko Polak. (vrpolak, 14:02:51)
    2. Dave Wallace (dwallacelf, 14:02:53)

  2. Agenda bashing (mackonstan, 14:03:38)
  3. FD.io CSIT lab infrastructure (mackonstan, 14:06:47)
    1. new HW started to arrive (mackonstan, 14:07:04)
    2. agreed on csit-dev to go with 2n testbeds (mackonstan, 14:07:16)

  4. Inputs from LFN and FD.io projects - VPP (mackonstan, 14:07:57)
    1. Dave: 19.01.3 maintenance releasing today, 19.04.2 next Wed. (mackonstan, 14:08:23)
    2. Jan Gelety (jgelety, 14:08:26)
    3. CSIT usual maintenance practice (mackonstan, 14:11:13)

  5. HoneyComb (mackonstan, 14:11:39)
    1. no updte (mackonstan, 14:11:42)

  6. Sweetcomb (mackonstan, 14:11:50)
    1. no update (mackonstan, 14:11:54)

  7. TSC (mackonstan, 14:11:57)
    1. Maciek: some budget available for spares (mackonstan, 14:12:11)
    2. TRIGGER 1901.2 PERF NDRPDR SEL1 (jgelety, 14:13:34)
    3. https://gerrit.fd.io/r/19740 (jgelety, 14:13:34)
    4. Maciek: discussion about FD.io presence at events, e.g. ONS EU, Kubecon. No budget/ROI for FD.io minisummits (costs too high), but suggested informal FD.io meetups for interested parties, advertised by LFN/FD.io, this is being pursued by FD.io mktg team / Trishan (mackonstan, 14:13:47)

  8. back to testing maintenance releases (mackonstan, 14:16:23)
    1. Maciek: proposal, vpp code changes only in feature, so testing every driver combination is pointless. (mackonstan, 14:17:10)
    2. Maciek: proposal - x710, 2n-skx, 3n-skx ipsec-sw only, 3n-hsw ipsec-hw only (mackonstan, 14:19:39)
    3. sounds good (mackonstan, 14:19:57)
    4. AGREED: (mackonstan, 14:20:01)
    5. AGREED: (mackonstan, 14:20:09)

  9. Operational status (mackonstan, 14:20:52)
    1. Vratko: vpp-csit-verify tests failing due to inability to compile vpp (mackonstan, 14:21:23)
    2. Peter: out of memory error while compiling (mackonstan, 14:26:46)
    3. Ed: to follow thru and report back (mackonstan, 14:28:01)
    4. https://jenkins.fd.io/view/vpp/job/vpp-csit-verify-device-master-1n-skx/2274/console (mackonstan, 14:29:21)
    5. something changed that made the compilation fail (mackonstan, 14:29:34)
    6. Ed: vpp-csit-verify jobs are using unique build sequence (mackonstan, 14:31:05)
    7. Ed: adjusting build script in ci-mgmt to t/shoot this problem (mackonstan, 14:31:24)
    8. from above link: ccache: error: Failed to fork: Cannot allocate memory (mackonstan, 14:31:58)
    9. some trending job tests failing due to infra (mackonstan, 14:33:23)
    10. ignore above (mackonstan, 14:33:34)
    11. some daily mrr jobs failiing due to new PAPI infra (mackonstan, 14:34:00)
    12. some daily mrr jobs tests failiing due to new PAPI infra (mackonstan, 14:34:54)
    13. https://gerrit.fd.io/r/20170 (mackonstan, 14:35:09)
    14. fixed by (mackonstan, 14:35:13)
    15. https://gerrit.fd.io/r/19272 (mackonstan, 14:35:30)
    16. another issue with PAL is fixed by this patch (mackonstan, 14:35:43)
    17. https://gerrit.fd.io/r/20695 (mackonstan, 14:35:54)

  10. Centos CI (mackonstan, 14:36:24)
    1. Ed: did few rounds with Centos image, more testing needed (mackonstan, 14:36:57)

  11. HoneyComb tests (mackonstan, 14:37:06)
    1. Vratko: reporting that HC tests are not working, as HC code is not keeping up to speed with VPP changes on master (mackonstan, 14:37:53)
    2. Vratko: propose to either remove the unused KWs or produce new tests that make use of them (mackonstan, 14:41:12)
    3. on csit-dev pls provide the list of KWs involved and related VPP functionality tested on VPP, and then we as a project team can make a call (mackonstan, 14:43:24)

  12. API test coverage (mackonstan, 14:43:40)
    1. Vratko: working on API CRC checks, and discovering which APIs are used by performance tests (mackonstan, 14:44:29)
    2. Vratko: will provide the list of APIs used in performance and device (mackonstan, 14:46:02)
    3. then we can see what the delta is (mackonstan, 14:47:23)

  13. VPP code performance (mackonstan, 14:47:32)
    1. no update, as tests are failing (mackonstan, 14:47:44)
    2. Vratko: next week results will be there for this week (mackonstan, 14:48:09)

  14. VAT to PAPI (mackonstan, 14:48:49)
    1. Vratko: big change merged - remote socket PAPI https://gerrit.fd.io/r/19272 (mackonstan, 14:49:09)
    2. Vratko: all PAPI calls go via remote socket apart from VPPstats (mackonstan, 14:49:37)
    3. Vratko: moving now to csit vpp api gating based on crc checks (mackonstan, 14:50:20)
    4. Jan: IPsecutil merged, srv6util and lisputil wip (mackonstan, 14:51:25)

  15. New RF tags proposal (mackonstan, 14:58:13)
    1. Maciek: presented the latest draft version of RF tags proposal (mackonstan, 14:59:15)
    2. in markdown format: (mackonstan, 14:59:53)
    3. https://drive.google.com/open?id=1n74mmbA2HSX-M4hBs3S39OA_KMSBFzKq (mackonstan, 14:59:56)

  16. other biz (mackonstan, 15:00:09)
    1. Ed: looks like the vpp-csit-verify issues are due to gcc8 (mackonstan, 15:00:29)
    2. Dave: taking action (mackonstan, 15:00:35)


Meeting ended at 15:01:06 UTC (full logs).

Action items

  1. (none)


People present (lines said)

  1. mackonstan (67)
  2. collabot` (4)
  3. jgelety (3)
  4. vrpolak (1)
  5. dwallacelf (1)


Generated by MeetBot 0.1.4.