Odd behaviour after upgrade to R80.20

[5lumber]

Hi,

We have a strange behavior we observing after upgrading from R77.20 standalone to R80.20 distributed.

Production Box: R77.20 standalone

Loan unit: R80.20 distributed.

We trying to swap the standalone from R77.20 to R80.20 but whenever we tried to to plug in the interfaces on R80.20 loan unit, the connection from gateway to next adjacent core switch went haywire.

The fw is supposed to connect to core switch on eth2

eth2 Link encap:Ethernet HWaddr 00:1C:7F:3D:C9:DD
inet addr:172.31.190.1 Bcast:172.31.190.3 Mask:255.255.255.252

As you can see it only has two IP, the next one is the core switch IP, 172.31.190.2

Whenever the external interface is connected to the box, the ping to the core switch went about 16-18 hops and tracert comes back with huge latency.

We tried almost anything we can think off but still couldn't found the answers, we tried spin another standalone box with same R77.20 and we face the same issue again.

Any help ?

[PhoneBoy]

Sounds like an issue with the switch to me.
Or a layer 1/2 issue.
What do the interface statistics say?