#fdio-csit: 12-Jul FD.io CSIT Project Meeting
Meeting started by mackonstan1 at 14:02:41 UTC
(full logs).
Meeting summary
-
- Dave Wallace (dwallacelf,
14:03:02)
- Ed Kern (snergster,
14:03:30)
- Inputs from LF and FD.io projects (mackonstan1, 14:03:32)
- Peter Mikus (pmikus1,
14:03:33)
- (mackonstan1, 14:03:36)
- Jan Gelety (jgelety,
14:03:36)
- operational status (mackonstan1, 14:06:57)
- vpp project reported: 43 rechecks since last on
6/20 (mackonstan1,
14:08:39)
- unclear if any are due to infra or csit test
issues (mackonstan1,
14:08:51)
- Peter will check if any performance regression
based on trending jobs (mackonstan1,
14:09:38)
- VIRL ssh timeout issues for tb4-virl1 and tb4-virl2 [FD.io Helpdesk #41921] (mackonstan1, 14:10:07)
- Ed reporting - running pybot locally on virl
controller node on tb4-virl2 => on ubuntu tb-virl2: no critical
tests failing, 20 complete suite runs. (mackonstan1,
14:15:38)
- Ed reporting - this verifies the integrity of
the local virl environment itself, excluding IP connectivity with
outside. (mackonstan1,
14:16:27)
- Ed reporting - on ubuntu: some non-critical
tests are failiing, as expected. (mackonstan1,
14:17:40)
- Ed reporting - on centos: some critical tests
are failiing, LISP and VM, as expected. (mackonstan1,
14:17:54)
- Ed reporting - next steps: 1. continue to
execute to clean up Centos, 1 or 2 failures. (mackonstan1,
14:22:26)
- Ed reporting - next steps: 2. put virl control
node on vilr1, and run against virl2, to get clean 20 complete suite
runs, ubuntu only. (mackonstan1,
14:23:02)
- Ed reporting - next steps: 3. talk to Vanessa,
ask for place to spin manually jenkins slave to run jobs with
simulations and tests on tb4-virl1 and tb4-virl2. (mackonstan1,
14:24:01)
- Tom Herbert (tfherbert,
14:25:07)
- Ed reporting - 25 complete runs, and clean for
ubuntu. (mackonstan1,
14:27:15)
- un-pegging dpdk dev package version from CSIT (mackonstan1, 14:27:39)
- another patch identified based on failures of
updated gerrit patch from Sergio (mackonstan1,
14:29:10)
- https://gerrit.fd.io/r/#/c/7529/
(mackonstan1,
14:29:13)
- rls1707 plan (mackonstan1, 14:29:46)
- Stateful Security Groups - InTesting - code
complete, tests running, 2t2c ip4 path, N aces reflect, 100k flows
=> VPP crashes. (mackonstan1,
14:32:52)
- NSH_SFC - functional - 24 tests, 1
failiing. (mackonstan1,
14:34:45)
- VM vhost use cases - InTesting - tb3 testbed
"misbehaving" due to wrong irq loadbalancing. tb1 and tn2 yielding
high stdev values for 2t2c tests. (mackonstan1,
14:40:10)
- DPDK17.05 testpmd - Done - few tests added,
i40e fixes in code and configuration. (mackonstan1,
14:42:01)
- Centos7 VIRL tests - WIP - Tom keeps working on
getting the simulations working, getting qemu hangs. using qemu
2.6. (mackonstan1,
14:45:18)
- LXC memif use cases - InTesting - vpp crashes
under load fixed, another patch from Damjan, for fixing spotted cpu
stalls. (mackonstan1,
14:49:05)
- l2fwd-cryptodev baseline - will not make it due
to cryptodev library update issues (mackonstan1,
14:53:31)
- adding test equipment: Netgate (mackonstan1, 14:53:56)
- Tom updating - documented the upgrade
procedure, will execute upgrade today. (mackonstan1,
14:54:44)
- next step is to complete the playbooks,
topology files, and start onboarding (mackonstan1,
14:55:15)
- rls1704 report (mackonstan1, 15:01:25)
- report is auto-generated, but some sections,
notably rls notes are handcrafted (mackonstan1,
15:01:52)
- Peter and Maciek reviewed the latest status,
and identified all work needed for completing generation of rls1707
report (mackonstan1,
15:02:42)
- Maciek will file jira tickets for all of these,
so that folks can grab the items for editing and reviewing, and we
can track progress (mackonstan1,
15:03:19)
- aiming for complete draft version to be
generated into master by end of this week (mackonstan1,
15:03:58)
Meeting ended at 15:04:01 UTC
(full logs).
Action items
- (none)
People present (lines said)
- mackonstan1 (37)
- collabot` (4)
- tfherbert (1)
- jgelety (1)
- dwallacelf (1)
- snergster (1)
- pmikus1 (1)
Generated by MeetBot 0.1.4.