#fdio-meeting: FD.io CSIT Project Meeting
Meeting started by mackonstan at 14:04:11 UTC
(full logs).
Meeting summary
-
- Vrato Polak. (vrpolak,
14:07:10)
- mackonstan (mackonstan,
14:07:14)
- Agenda bashing (mackonstan, 14:07:29)
- VPP update (mackonstan, 14:08:10)
- TSC (mackonstan, 14:08:22)
- Dave - LFN dev &test forum next week -
three fd.io talks: csit dashboard thu 6-7am pdt, sonic-vpp intro thu
5-6am pdt, strongswan-vpp wed 5-6am pdt (mackonstan,
14:09:37)
- https://wiki.lfnetworking.org/display/LN/2023-06+-+FD.io%3A+StrongSwan-VPP+plugin+Integration+Testing
(5am PDT, Wednesday June 7, 2023) (mackonstan,
14:11:24)
- https://wiki.lfnetworking.org/display/LN/2023-06+-+FD.io%3A+SONiC+VPP+platform+Use+Case
(5am PDT, Thursday June 8, 2023) (mackonstan,
14:11:30)
- https://wiki.lfnetworking.org/display/LN/2023-06+-+FD.io%3A+CSIT+Performance+Dashboard
(6am PDT, Thursday June 8, 2023) (mackonstan,
14:11:36)
- CSIT-2302 Performance RCAs (mackonstan, 14:14:00)
- Vratko - re CSIT-1902 3n-alt: Changed
performance in VXLAN and IPsec tests - looking at trending, the
behaviour changed again, seeing both good and bad
performance. (mackonstan,
14:15:05)
- Vratko - re RSS is not deterministic on
Intel-E822CQ NIC - waiting for vpp patch (mackonstan,
14:15:43)
- old patch (mackonstan,
14:17:12)
- 38304: interface dpdk avf: introducing setting RSS hash key
feature | https://gerrit.fd.io/r/c/vpp/+/38304
(mackonstan,
14:17:14)
- ACTION: Kai to check
with Intel contributors re what's happening with this patch
(mackonstan,
14:19:56)
- CSIT-2306 (mackonstan, 14:20:25)
- mackonstan on behalf of Tibor: rc1 tests are in
progress - iterative tests done (mackonstan,
14:21:37)
- Vratko - same failures as present in
trending (mackonstan,
14:22:04)
- Physical Infrastructure (mackonstan, 14:22:40)
- Dave re SPR and ICX-D orders - still waiting
for updates from Intel re CPU and NIC donations blocking SuperMicro
and Vexxhost quotes (mackonstan,
14:25:53)
- Dave re NICs and cables needed for the new 4
SPR systems to be ordered - Peter can use spare NICs left over from
decommissioned SKXs (mackonstan,
14:30:51)
- ACTION: mackonstan
and Peter to work through existing NIC inventory, incl. FVL, CVL and
cx6/cx7, and report back if any more NICs and cables needed
(mackonstan,
14:31:35)
- ACTION: Dave to watch
us (mackonstan,
14:31:41)
- mackonstan on behalf of Peter: 2n-spr working
in trending and rc1 of 23.06 (mackonstan,
14:32:12)
- mackonstan on behalf of Peter: 3nb-spr with CVL
e810cq nics - all tests are working (mackonstan,
14:33:25)
- mackonstan on behalf of Peter: 3na-spr with cx7
2p200ge nics - 80% of single link tests are working, intermittent
trex cx7 initialization failures that are being t/shooted
(mackonstan,
14:34:11)
- Virtual Infrastructure, CI/CD (mackonstan, 14:35:32)
- vpp-csit-verify-device jobs are back and
voting (mackonstan,
14:35:51)
- Trending and Test Failures (mackonstan, 14:36:07)
- Vratko - no changes, status in wiki
(mackonstan,
14:36:41)
- Vratko - bump of TRex to v3.0.3 is merged but
not in oper branch yet (mackonstan,
14:37:23)
Meeting ended at 14:37:55 UTC
(full logs).
Action items
- Kai to check with Intel contributors re what's happening with this patch
- mackonstan and Peter to work through existing NIC inventory, incl. FVL, CVL and cx6/cx7, and report back if any more NICs and cables needed
- Dave to watch us
Action items, by person
- mackonstan
- mackonstan and Peter to work through existing NIC inventory, incl. FVL, CVL and cx6/cx7, and report back if any more NICs and cables needed
- UNASSIGNED
- Kai to check with Intel contributors re what's happening with this patch
- Dave to watch us
People present (lines said)
- mackonstan (33)
- collab-meetbot (9)
- vrpolak (1)
- dwallacelf (0)
- jlinkes (0)
- tifrank (0)
Generated by MeetBot 0.1.4.