#fdio-vpp: fdio-vpp
Meeting started by DaveBarach at 15:01:34 UTC
(full logs).
Meeting summary
-
- Daily perf tests - failures: (mackonstan,
15:03:36)
- daily perf runs throttled back due to 19.08
CSIT report generation / testbed capacity (DaveBarach,
15:03:53)
- - papi dump calls, remote socket connection not
opening, intermittent, affects ~3% of tests (mackonstan,
15:03:59)
- - memif activation failures,
intermittent (mackonstan,
15:04:12)
- - avf activation failures, intermittent
(mackonstan,
15:04:23)
- CSIT-1908 release status (mackonstan,
15:04:34)
- Draft version of CSIT-1908 test and
benchmarking report is available on the CSIT master report
link (mackonstan,
15:04:45)
- https://docs.fd.io/csit/master/report/index.html
(mackonstan,
15:04:58)
- Most tests on Xeon, Atom and Arm testbed are
completed, some are finishing tomorrow (mackonstan,
15:05:19)
- Some VPP throughput regressions and
progressions spotted - see my email to vpp-dev and csit-dev
yesterday (mackonstan,
15:05:32)
- Some are CSIT, some look like VPP, see latest
analysis in CSIT gerrit commits: (mackonstan,
15:05:54)
- - csit1908_perf_analysis_2n_skx.md:
https://gerrit.fd.io/r/c/csit/+/21637/ (mackonstan,
15:06:01)
- - csit1908_perf_analysis_3n_skx.md:
https://gerrit.fd.io/r/c/csit/+/21641/ (mackonstan,
15:06:09)
- "Official" CSIT-1908 report will be published
tomorrow, on schedule :) (mackonstan,
15:06:16)
- https://docs.fd.io/csit/rls1908/report/index.html
(mackonstan,
15:06:21)
- Will send announce email per usual (mackonstan,
15:06:27)
- Host Stack (Florin reporting) (DaveBarach, 15:07:15)
- moving control API calls from binary API to the
host stack message queue (DaveBarach,
15:07:32)
- old APIs are still operational, planning to
remove them post 20.01 (DaveBarach,
15:08:00)
- async / non-blocking connect added for
Envoy (DaveBarach,
15:08:22)
- TCP improved mss handling for v4, v6
(DaveBarach,
15:08:34)
- if anyone is interested in host-stack crypto,
Nathan has been working on it for TLS (DaveBarach,
15:09:05)
- simon @ intel has a TSO patch in progress
-major vector rate reduction for TX side (DaveBarach,
15:09:58)
- Documentation (John, Ray) (DaveBarach, 15:13:36)
- RX marketing section feeback from Ray
(DaveBarach,
15:13:50)
- ACTION: jdenisco to
try to pick a cut-over date (DaveBarach,
15:16:17)
- aarch64 (Honnapa reporting) (DaveBarach, 15:17:23)
- dual-loop optimization, bihash optimization in
progress (DaveBarach,
15:17:41)
- 19.04 (Dave Wallace) (DaveBarach, 15:18:04)
- few patches no formal release planned at this
time (DaveBarach,
15:18:25)
- 19.08 (Andrew reporting) (DaveBarach, 15:18:46)
- 8 commits merged since formal release
(DaveBarach,
15:18:56)
- possible change to api string type (DaveBarach,
15:20:04)
- question: can we merge it (DaveBarach,
15:20:09)
- ACTION: ole hook up
with ray to discuss the vppapigen VLA issue (DaveBarach,
15:26:48)
- "while we're at it" a couple of memif API
issues could be resolved (DaveBarach,
15:27:48)
- zero-copy control disappeared, could fix
it (DaveBarach,
15:28:29)
- API changes (DaveBarach, 15:33:35)
- jerome: about to release vpp 19.08
openstack (DaveBarach,
15:34:00)
- acl add/replace had a CRC change - name of type
of acl rule type caused the change (DaveBarach,
15:35:39)
- https://gerrit.fd.io/r/q/project:vpp+is:open+path:%5E.*/.api
(dwallacelf,
15:46:26)
- Please ignore previous link -- it doesn't work
with latest gerrit version. (dwallacelf,
15:47:37)
- infra (Ed K) (DaveBarach, 15:49:50)
- "runtest" = "recheck," (DaveBarach,
15:50:01)
Meeting ended at 15:52:16 UTC
(full logs).
Action items
- jdenisco to try to pick a cut-over date
- ole hook up with ray to discuss the vppapigen VLA issue
People present (lines said)
- DaveBarach (30)
- mackonstan (16)
- collabot` (3)
- dwallacelf (2)
- mortderire (1)
Generated by MeetBot 0.1.4.