Arp can't be sent out by Trex(V2.96)

[liuguili]

Hi,
Using Trex v2.96 with two ports connected to DCGW, I can see Trex port0 and port1 try to send arp request, but didn't send out finally, I suspect there is something wrong with the bottom layer.
logs:
TX ARP request on port 0 - ip: 11.108.4.1 vlan: 467 mac: Unknown
+---------+---------------+----------+
00:00:00,000,000 ETHER
|0 |ff|ff|ff|ff|ff|ff|72|ba|e2|2c|83|44|81|00|01|d3|08|06|00|01|08|00|06|04|00|01|72|ba|e2|2c|83|44|0b|6c|04|58|01|03|05|07|09|00|0b|6c|04|01|00|00|00|00|00|00|00|00|00|00|00|00|00|00|

TX ARP request on port 1 - ip: 11.108.5.1 vlan: 468 mac: Unknown
+---------+---------------+----------+
00:00:00,000,000 ETHER
|0 |ff|ff|ff|ff|ff|ff|d6|bf|ba|eb|0c|14|81|00|01|d4|08|06|00|01|08|00|06|04|00|01|d6|bf|ba|eb|0c|14|0b|6c|05|58|01|03|05|07|09|01|0b|6c|05|01|00|00|00|00|00|00|00|00|00|00|00|00|00|00|

oerrors == 11, proof that something went wrong during the sending process.

-Per port stats table
ports | 0 | 1

---

opackets | 0 | 0
obytes | 0 | 0
ipackets | 10 | 10
ibytes | 680 | 680
ierrors | 0 | 0
oerrors | 11 | 11
Tx Bw | 0.00 bps | 0.00 bps

Trex is deployed to a pod in Kubernetes, issue occures after the test channel was re-orchestrate in dpdk mode, and is fixed if we delete Trex pod and it never pops up for later runs on the same channel.

[liuguili]

Below is complete failed log:

failed_trex.log