#fdio-csit: FD.io CSIT project meeting
Meeting started by mackonstan at 14:01:51 UTC
(full logs).
Meeting summary
- Agenda bashing (mackonstan, 14:04:50)
- Tom Herbert (tfherbert,
14:05:04)
- Dave Wallace (dwallacelf,
14:05:22)
- FD.io CSIT lab infrastructure (mackonstan, 14:05:50)
- skip due to Mohammed not present (mackonstan,
14:06:01)
- Inputs from LFN and FD.io projects (mackonstan, 14:06:23)
- Peter Mikus (pmikus,
14:06:44)
- VPP update - DaveW - few more patches in stable/1904 (mackonstan, 14:06:51)
- DaveW - few more patches in stable/1904
(mackonstan,
14:07:02)
- DaveW - some issues with Arm merge jobs
(mackonstan,
14:07:39)
- HC (mackonstan, 14:07:42)
- no updates (mackonstan,
14:07:47)
- TSC (mackonstan, 14:07:59)
- Maciek - escalated to TSC two issues: i)
jenkins jnlp4 connection drops; ii) denverton testbed access issues,
preventing/delaying CSIT-1904 release. (mackonstan,
14:10:06)
- issue i) addressed by removing ha-proxy from
the path; ii) addressed by correct network infrastructure
configuration. (mackonstan,
14:10:48)
- Operational status (mackonstan, 14:11:38)
- EdK - few fires, but nothing CSIT specific
(Centos image related) (mackonstan,
14:12:41)
- Centos CI testing status - Tom - image updated,
patch merged. Will work on testing the image in the VPP_device
testbeds. (mackonstan,
14:14:28)
- VPP code performance (mackonstan, 14:14:48)
- trending graphs looking okay, but testing gaps
due to jnlp4 induced outages (mackonstan,
14:16:30)
- release CSIT-1904 (mackonstan, 14:16:52)
- report data collection testing ongoing
(mackonstan,
14:18:18)
- tests still missing: 3n-hsw, nfvdensity, soak
tests (mackonstan,
14:21:55)
- issues: ipsec tunnel-interfaces with
aes-128-gcm - low performance, failures are still under
investigation (mackonstan,
14:23:13)
- very sporadic issues with avf initialization
and with memif - CSIT tickets open, rls notes to be updated
(mackonstan,
14:24:36)
- due to testing being delayed, suggestion to
delay the report by 2 days to 10-May (mackonstan,
14:25:35)
- Developments (mackonstan, 14:26:00)
- VAT to PAPI migration - in progress
(mackonstan,
14:27:40)
- PAPI execution over the remote UNIX socket -
Ole and Vratko experimenting (mackonstan,
14:28:09)
- need to make a decision if this is the right
way forward - see #csit-papi FD.io slack channel (mackonstan,
14:28:48)
- automate VPP-API "Flag Day" - DaveW - editing
almost-final version - see patch (mackonstan,
14:32:07)
- https://gerrit.fd.io/r/#/c/18356/
(mackonstan,
14:32:15)
- VPP perfmon plugin - Peter - APIs missing,
stats are not exported into the stats segment, to be addressed with
vpp-dev/Ole (mackonstan,
14:34:00)
- New tests (mackonstan, 14:34:30)
- IPsec - Maciek, Peter - reworking IPsec tests
following the new VPP crypto developments (mackonstan,
14:35:09)
- TRex (mackonstan, 14:35:22)
- Peter - upgrade on hold till we're out of 1904
release (mackonstan,
14:35:34)
- VIRL servers repurpose(?) (mackonstan, 14:35:55)
- EdK - VIRL3 used for build infra (mackonstan,
14:36:08)
- current users of VIRL: HC, DMM (mackonstan,
14:37:12)
- Fd.io Events (mackonstan, 14:43:47)
- https://fdiominisummiteu19.sched.com/
(mackonstan,
14:46:12)
- vpp_device tests (mackonstan, 14:48:26)
- DaveW - shouldn't we enable voting for
vpp-csit-verifyvpp-csit-verify-device-master-1n-skx (mackonstan,
14:49:44)
- Peter - will push the patch to ci-mgmt to
address this (mackonstan,
14:52:07)
- Peter - more tests need to be added to
vpp_device: avf driver, multiple rxq tests for dpdk, avf,
virtio/vhost, memif (mackonstan,
14:53:36)
Meeting ended at 14:53:42 UTC
(full logs).
Action items
- (none)
People present (lines said)
- mackonstan (45)
- collabot (3)
- tfherbert (1)
- dwallacelf (1)
- pmikus (1)
Generated by MeetBot 0.1.4.