We have a CX320 that is connected to two file servers one 2650, and one 2950 PE server.  Both use QLogic HBAs, of course the 2650 unit is PCI-X, and the 2950 is PCI-E.  We rolled out to users on Monday and the 2650 has hung up 4 times in the 3 days since load was put on the servers.  The SAN fabric is two Cisco switches and both servers have a pair of HBAs, each connecting to opposite but corresponding SPs on the fabric, (i.e., 1 switch managing SPa-0 and SPB-1, 2 switch managing SPa-1 and SPB-0) as Dell and EMC both recommend is best practice.  When we began troubleshooting this afternoon, after extensive searching through the event logs, and reboots to get the file shares back online.  We found that Navisphere showed the 2650 as an unknown host, marked with a U in the GUI.  So we double checked the navi agent config files on the server and they appear to be correct.  We then removed the luns & storage group from the server and unregistered all in navisphere, then restarted the server to get the host to remove from navisphere on the storage tab, it never disspeared from the host tab.  Then we removed navi CLI, navi Agent, Powerpath, SANsurfer from the server and restarted.  Removed the fibre cables and restarted the server, and started adding the software back in this order - SANsurfer first, making sure to run the registration wizard for both HBAs then we connected the fibre cables back, then installed Powerpath, and then NAVi agent -working through the registration wizard, and then navi cli.  After all this, the navisphere website brought back up the server in the storage tab and it came back with a U, when you check properties, it says the OS is "unkown", this server is running Windows Server 2003 Enterprise R2, with SP2 installed, just as the 2950.  It seems that no matter what we try we can NOT remove the "U" marker in navisphere, and we believe that is the root of the server hangs.  There have not been any errors in the windows event logs prior to or directly after the hang ups.  We have re-verified DNS, and host resolution b/w the server and the CX320 SPs, all resolve.  We have set up dependencies with the networker remote execute service and the ONC/Portmap service so as not to cause conflicts, but still we have these hangups.  We moved the print spool from the system drive to a data drive - not SAN based and still the file shares hosted on the SAN LUNS on this server become unreachable.  We are at a loss to determine what else could be going on.  Any suggestions to resolve the "U" problem or to track down the file serviing hang ups would be appreciated.