Comware

I have just deployed 200+ H3C 3100 v1 switches in a mixed Cisco environment. No issues.

We tried to order an additional 50+ H3C switches, but were told we could no longer get them.

So, HP shipped us 50 new HP 3100 v2 switches instead. No problem. Com3 vs. Com5

Only, now we are randomly losing ALL connectivity on the v2 switch ONLY.

TAC has been looking into the problem for over a week now with no direction.

We thought is was possibly a STP problem - MSTP vs. PVST. So, we turned off STP.

We still lost ALL connectivity, but only on the v2 switches again. ALL v1 were fine.

Before turning off STP, it appeared like the v2 were ignoring the PVST packets and possibly

sending MSTP to each other, causing a topology update/change to each vlan instance. But...

Once we disabled STP and it still happened... (?)

Additionally, if we wipe the config and connect the v2 switches to an access port, they function fine.

TAC has verified our baseline config on both the v1 and v2 switches along with our Cisco uplink ports.

So, I figured I'd jump onto the forums and see if anyone else out there is having a similar issue...

Thanks for your immediate response...

Well, it appears to be a bug in the 5103 code. We are running:

Comware Software, Version 5.20, Release 5103P01

This version has some type of ARP problem whereas it

flushes the cache once it receives more than 200 packets/sec.

A trunked link with multiple vlan instances will definitely push that...

So, we are going to load H3C Asian Version Release 5106 and see if the problem goes away.