Trex counters [0] see one direction loses all packets, but VPP counters [1] see all packets present and no issues. Other testbeds have no issue with jumbo+geneve tests, so I am not sure whether this is invisible TX failure on VPP side (MTU is large enough) or on TRex RX side. More investigation is needed.
[0] https://logs.fd.io/vex-yul-rot-jenkins-1/csit-vpp-perf-report-coverage-2606-2n-zn2/28631712876/log.html.gz#s1-s1-s1-s1-s2-t6-k3-k7-k1-k1-k1-k8-k13-k1-k2
[1] https://logs.fd.io/vex-yul-rot-jenkins-1/csit-vpp-perf-report-coverage-2606-2n-zn2/28631712876/log.html.gz#s1-s1-s1-s1-s2-t6-k3-k7-k1-k1-k1-k8-k14-k1-k1-k1-k1
Trex counters [0] see one direction loses all packets, but VPP counters [1] see all packets present and no issues. Other testbeds have no issue with jumbo+geneve tests, so I am not sure whether this is invisible TX failure on VPP side (MTU is large enough) or on TRex RX side. More investigation is needed.
[0] https://logs.fd.io/vex-yul-rot-jenkins-1/csit-vpp-perf-report-coverage-2606-2n-zn2/28631712876/log.html.gz#s1-s1-s1-s1-s2-t6-k3-k7-k1-k1-k1-k8-k13-k1-k2
[1] https://logs.fd.io/vex-yul-rot-jenkins-1/csit-vpp-perf-report-coverage-2606-2n-zn2/28631712876/log.html.gz#s1-s1-s1-s1-s2-t6-k3-k7-k1-k1-k1-k8-k14-k1-k1-k1-k1