14:57:49 <mackonstan> #startmeeting FD.io CSIT project meeting 14:57:49 <collabot`> Meeting started Wed Feb 6 14:57:49 2019 UTC. The chair is mackonstan. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:57:49 <collabot`> Useful Commands: #action #agreed #help #info #idea #link #topic. 14:57:49 <collabot`> The meeting name has been set to 'fd_io_csit_project_meeting' 14:58:05 <mackonstan> #topic Agenda bashing 14:58:13 <mackonstan> #chair 14:58:13 <collabot`> Current chairs: mackonstan 15:01:16 <jgelety> #info Jan Gelety 15:02:41 <mackonstan> #topic FD.io CSIT lab infrastructure 15:03:18 <mackonstan> #info no open issues on infra, so moving swiftly on 15:05:11 <mackonstan> #info Mohammed mgmt switch with qsfp interfaces is installed, to be used for managing thunderx2 servers 15:06:10 <mackonstan> #info Mohammed wiring installed per specs provided by Arm team, ports are not up as switch needs to be configured 15:07:03 <mackonstan> #info Sirshak correction: above mgmt switch is for thunderx1's not for thunderx2's. (latter is managed by SFP links) 15:08:23 <mackonstan> #info Juraj - QSFP cables for thunderx2 servers shipped, tracking info provided. 15:08:58 <mackonstan> #topic Inputs from LFN and FD.io projects 15:11:17 <mackonstan> #info VPP project - Maciek reporting - vpp-19.01 shipped, Andrew will write up suggestions to improve release process 15:12:27 <mackonstan> #info VPP project - Maciek - after buffer code rewrite, Damjan is reporting +25% higher MRR throughput for AVF paths in his testbed vs. CSIT trending. Need to investigate. 15:13:56 <mackonstan> #info HC - Michal reporting - HC 19.01 shipped, CSIT tests passing with 2 failures expected, fix coming into master, broken test, not HC code 15:14:59 <mackonstan> #topic Operational status 15:15:59 <michal-cmarada> #info HC - CSIT known issue VPP-649 for honeycomb resolved, tests will be fixed and reenabled in 1904 15:16:02 <mackonstan> #info no issues 15:16:06 <mackonstan> #topic VPP code performance 15:17:37 <mackonstan> #info Jan reporting: vpp failures - ipsec aes-gcm fixed now, issue was with activating dpdk crypto backend. 15:18:28 <mackonstan> #info Jan: vhost vhost/vm failures - issue with 15:19:02 <mackonstan> #info Jan: vhost vhost/vm failures - issue with VM install getting corrupted, workaround applied. 15:19:36 <mackonstan> #info Jan: vhost/vm 4c tests still failing, debugging with Peter 15:20:28 <mackonstan> #info Jan: lb tests - not reproducible by manual tests in CSIT testbeds, only happens with Jenkins driven tests, implementing coredump collection to catch the guy 15:22:40 <mackonstan> #info Peter: avf multi-core tests failing now too.. 15:22:58 <mackonstan> #topic Releases 15:24:26 <mackonstan> #info Tibor: most static content is complete. Methodology section updated and split into multiple pages due to growth in size. 15:25:17 <mackonstan> #info Tibor: added graphs: i) NFV service density tests, ii) soak tests. 15:25:28 <mackonstan> #link https://docs.fd.io/csit/master/report/vpp_performance_tests/nf_service_density/vnf_service_chains.html 15:27:28 <mackonstan> #info Tibor: need to update rls notes for all sections, dmm/nsh based on working tests, HC sections, waiting for 1901 test results data 15:32:27 <mackonstan> #info 1901 test data, Jan: delay due to failing tests ( lb, vhost, memif). Selected performance: 3 runs finished, 3 running, remaining 4, expecting all to complete by Monday if no other issues. 15:33:37 <mackonstan> #info above line was for 2n-skx. 3n-skx: 2 run finished, 3 running. 3n-hsw: 1 finished, 3 running. 15:35:19 <mackonstan> #info NFV density - Peter: all code done, tests work well, need to run them to collect data for report. 15:36:26 <mackonstan> #info soak test PLRsearch - Vratko: identified one issue that sporadically prevents convergence.. working on a fix 15:36:56 <mackonstan> #topic Developments 15:38:09 <mackonstan> #topic framework 15:38:45 <mackonstan> #info Peter working on few improvements in VM and Container areas 15:41:17 <mackonstan> #info VAT to PAPI move - Peter assessment 1.5 man-months.. 15:41:44 <mackonstan> #topic Test environments 15:42:24 <mackonstan> #info Peter: preparing VPP job for running vpp_device jobs 15:42:47 <mackonstan> #info expecting to productize after csit-1901 rls cycles 15:43:29 <mackonstan> #topic TRex developments status 15:44:49 <mackonstan> #info Peter: lots of fixes in TRex releases. We should upgrade TRex in csit-1904 cycle (csit is on v2.3.5, trex latest v2.5.2) 15:45:18 <pmikus> #info https://github.com/cisco-system-traffic-generator/trex-core/blob/master/doc/release_notes.asciidoc 15:45:31 <mackonstan> #link https://github.com/cisco-system-traffic-generator/trex-core/blob/master/doc/release_notes.asciidoc 15:47:17 <mackonstan> #topic Housekeeping 15:48:37 <mackonstan> #info looking for volunteers to update the CSIT installation guide that's now outdated on both wiki and git repo 15:49:11 <mackonstan> #endmeeting