14:00:16 <mackonstan> #startmeeting FD.io CSIT Project Meeting
14:00:16 <collab-meetbot> Meeting started Wed Aug 26 14:00:16 2020 UTC and is due to finish in 60 minutes.  The chair is mackonstan. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:00:16 <collab-meetbot> Useful Commands: #action #agreed #help #info #idea #link #topic #startvote.
14:00:16 <collab-meetbot> The meeting name has been set to 'fd_io_csit_project_meeting'
14:00:29 <dwallacelf> #info Dave Wallace
14:01:25 <tifrank> #info Tibor Frank
14:03:22 <vrpolak> #info Vratko Polak.
14:03:30 <mackonstan> #topic Agenda bashing
14:03:51 <mackonstan> #topic Inputs from projects - VPP
14:04:31 <mackonstan> #info Dave, vpp v19.08.3 released yesterday.
14:13:53 <mackonstan> #info Dave, question is if v19.08.3 should be perf tested in CSIT, similar to what was done for 19.08.2 back in May.
14:14:38 <mackonstan> #info Maciek, CSIT project now busy with finish 20.09 code, quite a few CSIT infra and env changes …
14:15:22 <mackonstan> #info Dave, Andrew, Maciek, proposal to defer v19.08.3 testing till after 20.09, see if anybody having any issue with that approach
14:16:03 <mackonstan> #info Inputs from projects - VSAP
14:16:26 <mackonstan> #info quiet
14:16:36 <mackonstan> #topic Inputs from projects - TSc
14:17:17 <mackonstan> #info Vratko, it was a short call, the only topic was CSIT and terminology changes, nothing planned for 2009
14:17:58 <mackonstan> #info Vratko, VPP is already changing api's, but CSIT can continue using old ones
14:18:15 <mackonstan> #topic Release CSIT-2009
14:19:51 <mackonstan> #info Tibor, replacing old NAT44 deterministic tests (aka CGN) with new ones
14:20:13 <mackonstan> #topic Physical Infrastructure
14:22:27 <mackonstan> #info Peter, Xeon SSDs: 2 Nomad servers affected, drives replaced; 1 SSD RMAed; all SSDs' FWs upgraded by Peter, so not expecting any complaints in regards to SSDs dying due to old FW.
14:24:02 <mackonstan> #info Peter, re inventory: no SSDs spares left; 37 HDD spares are there.
14:26:12 <mackonstan> #info Peter, re Atom/Denverton, testbeds are still out-of-service, exhibit continuous pkt drops; similar setup in Intel labs doesn't suffer from this issue; all tuning settings are aligned between the setups; awaiting feedback from Intel experts re next steps.
14:29:21 <mackonstan> #info Juraj, re 2 faulty Arm servers: 1 Arm server recovered following BMC/IPMI config correction; 1 Arm server not booting due to incomplete RAM stick configuration, need to buy replacement RAM stick.
14:37:32 <mackonstan> #info Juraj, re new Arm testbed, 2x ThunderX2, 2x 2RU server, vexxhost confirmed no space in existing racks, need to get another rack, requires PO raised against
14:37:41 <mackonstan> #link https://secure.vexxhost.com/billing/viewticket.php?tid=FDU-009676&c=ncHht136
14:40:07 <mackonstan> #info Keesang, sending missing rack mounting kit
14:41:49 <mackonstan> #info Keesang, otherwise Vexxhost team is proceeding with the install
14:43:58 <mackonstan> #topic Virtual Infrastructure, CI/CD
14:45:00 <mackonstan> #info Peter, observed all Jenkins executors got suspended, gerrit triggers not taken care of. Andrew and Maros confirmed.
14:45:55 <mackonstan> #info Vanessa to be pinged.
14:46:14 <mackonstan> #info Vanessa on it per Andrew's update
14:47:06 <mackonstan> #info Andrew, making vpp-csit-verify-device voting, change uploaded, Vratko reviewed and merged.
14:48:00 <mackonstan> #into Dave, no TWS calls on infra last week and this week, will have one next week
14:48:31 <mackonstan> #topic Environment, Framework
14:51:04 <mackonstan> #info Vratko, TRex 3-ple patch squashed into 1 per call last week, merged into master, in oper branch next week, we should see the effects in trending.
14:53:25 <mackonstan> #info re VPP API changes' tracking, Maciek, been made aware of this change in stats api
14:53:29 <mackonstan> #link https://gerrit.fd.io/r/c/vpp/+/28017
14:53:51 <mackonstan> #info Andrew, this should be transparent to PAPI, but affects goVPP
14:54:14 <mackonstan> #topic Performance Tests
14:57:35 <mackonstan> #info Trending tests failing, based on csit-report auto email: AVF, RDMA, nat44
14:57:47 <mackonstan> #info Tibor, Peter, reporting some DPDK tests are failing too
14:58:23 <mackonstan> #info Vratko, no regressions in baseline tests, some progressions
15:00:37 <mackonstan> #info Vratko, natt44ed cps tests, (almost) ready for review, observing some TRex ASTF incosistency of behaviour
15:01:32 <mackonstan> #endmeeting