Hello all.
My first post here and wasn't quite sure should I use Networking or Server section. So here we go.

Old setup was:

Supermicro switch with 10Gb module where we had two servers connected. The other one is openindiana and it's used as a nfs server. The other one is a debian. Ssh and nfs worked just fine.

Now we need more 10Gb connections for the nfs.
I installed a new Cisco SG500XG switch and connected couple of new servers there plus the second one from the old switch. I also connected the Cisco to the old switch's 10Gb module. And the nfs server is still connected to that same port as before on that Supermicro switch.


And here's the tricky part:

New servers are able to connect via ssh and nfs to the nfs-server and ssh works just fine the other way round. But the old debian server (the one I unplugged from the old switch and connected to new Cisco) is not able to connect with the 10G interface to nfs-server anymore, but it can ssh to the rest of the servers in the same network. 1Gb works without any problems to nfs-server. And I'm able to ssh to that debian server from nfs-server, just not vice versa.

As told in Subject, telnet to those ports work. I've done googling etc. but found no help. "SSH2_MSG_KEXINIT sent" is the last thing I see before the connection hangs. Arp-table is up to date. Netmask is the same with every server. Any ideas? Thanks in advance.

With ssh it's the server side logs that usually reveal what went wrong.

I had to tamper with the MTUs to get this working.