Through packet captures using techniques described in KB 419729 - How to determine if IPFIX packets are flowing from DPU enabled ESXi hosts to VCF Operations for Networks Collector node(s):
ESXi
VCF Operations for Networks
The root cause is a code defect in the handling of gateway configurations for the "ops" netstack within the DPU architecture.
The gateway setting is not correctly propagating to the active routing table inside the DPU host.
NOTE: This is an ESXi code defect, but it impacts VCF Operations for Networks
The code fix for this defect will be included in a future release.
However, here is a Workaround:
esxcli network ip route ipv4 add -n 0.0.0.0/0 -g <GATEWAY_IP> -N ops"dpuctl ssh vmdpu0"esxcli network ip route ipv4 add -n 0.0.0.0/0 -g <GATEWAY_IP> -N ops"<GATEWAY_IP>" in the above commandvmkping -I vmkN ###.###.###.### -S ops"vmkN represents the vmkernel interface configured to the "ops" netstack###.###.###.### represents the IP address of the VCF Operations for Networks Collector.
If, after implementing the workaround, you still have difficulty ensuring that IPFIX packets leaving the DPU enabled ESXi host are being received by the VCF Operations for Networks Collector node, please open a support case with the VCF Operations for Networks team using the instructions at KB 142884 - Creating and managing Broadcom cases