Notes/Domino 6 and 7 Forum : SERIOUS problem with Domino after upgrading to 6.5 + changing hardware

Hi,
I'm having a seriuos problem with Domino after upgrading from 508 to 6.5. I've done this many times before and done it without any problems (following IBM guidelines aso). I'm doing this for a living as a consultant with many years of experience.

At a customer-site we upgraded a mail cluster and some application servers. There's only a problem with one of the servers, and that's the "main" server in the mail cluster.

We changed the hardware to some real serious stuff (IBM x345, Xeon 2,8GHz, 1,5GB RAM, OS on RAID1, Translog on RAID1 and data on RAID10 (EXP300 with 6 disks and RAID Controller 4MX)).

This server is the only one running W2K with SP4. All other servers running SP3.

There's 2 NIC's - one for LAN and one for Cluster.

Installed software:
W2K SP4
Domino 6.5
Symantec AV 3.1 build 78 (latest)
Domino tasks:
Router, SMTP
Replica
Cluster
Sched
AntiVirus
HTTP (minimal use)

The problem is; after some hours the server stop responding for the clients. On the console everything seems fine, mail routing works, replication works and so on. But the server can't find itself. Server "Server1/ACME" not responding and "Waiting tasks" (doing a sh server command) rises. After some minutes 40 - 60 waiting tasks is in queue.

Issuing the "sh cluster" command tells that the other server in the cluster is up and running, but the server itself is Unreachable.

I've checked the following:
The server is in the DNS
The server is named in the local HOSTS file
The server can ping itself, both by name and by IP-adress
The server name is correct in the "notes port" field (Server1) for port TCPIP
The server name is correct in the "notes port" field (172.16.1.1) for port CLUSTER
Issuing a "trace Server1/ACME" results in a connection using connection document (have a document pointing to itself just to try something out)

The following lines is in the notes.ini regarding the ports and cluster configuration:
Server_Cluster_On=1
Cluster_TcpConnectTimeOut=0,5
CLUSTER=TCP,0,15,0,,12288
TCPIP_TcpIpAddress=0,192.168.100.7:1352 (LAN connections)
CLUSTER_TcpIpAddress=1,172.16.1.1:1352 (Cluster connections)
Server_Cluster_Default_Port=CLUSTER
TCPIP_TcpConnectTimeOut=0,15
TCPIP=TCP, 0, 15, 0
Ports=TCPIP,CLUSTER

Any help will be appreciated very much, since this is a main mail server for hundreds of users.
TIA

Morten Schaumann