Conference decwet::winnt-clusters

609.0. "couldn't access shared if failover to the other server" by TLAV02::AKKAMON (Land of Smail :)) Sat Feb 08 1997 04:55

    Hi,
    
    	I am setting up the Alpha NT Cluster to show at the partner meeting
    this Monday morning.   I have a weird problem and have no idea how to
    fix it.
    	I installed the Cluster on Alpha 4100 5/300 and 2100 4/200 with
    KZPSA controllers, and named it "AlphaCluster".   The installation was 
    complete without any problem.  But when I start the Cluster Administrator 
    software on either server, I always get the message "Attempt to insert 
    duplicate SCSI Disk" which comes from "Target ID 0 - Logical Unit ID 0 - 
    DEC RZ28".
    	The shared SCSI bus was "SCSI port 3, bus 0, Initiator ID 6 and 7".
    I have 4 disks in the shared storage, BA356, targetted at Id 1,2,3,4.
    I fixed the drive letter to W, X, Y and Z respectively, then created
    4 failover groups as follow.
    		Drive W and Y  ----> belong to 4100 (nodename Rawhide)
    		Drive X and Z  ----> belong to 2100 (nodename Sable)
    	Then, created the shared directories in each shared disks.   
    		Drive W		---> W:\RawhideW
    		Drive X		---> X:\SableX
    		Drive Y		---> Y:\RawhideY
    		Drive Z		---> Z:\SableZ
    
    After that, I installed the Windows 95 with the Cluster client software. 
    Windows 95 can see the "Digital Cluster" in its "Network Neighborhood".
    I can get in Network Neighborhood and see the icon "AlphaCluster".  I
    then double click the AlphaCluster and it sees all the 4 shares, i.e.,
    RawhideW, RawhideY, SableX, and SableZ.
    	I can double click at either RawhideW and RawhideY without problem,
    I get the contents inside.   But I can't get into SableX or SableZ at
    all.   But if I access both SableX and SableZ thru the "Sable" not
    "Digital Cluster\AlphaCluster", it works!
    	I then try to do the failover of both SableX and SableZ so that the
    Rawhide takes over, the Windows 95 client can access them.   That means
    whenever the Sable controls the disks, clients cannot access at all.
    	What is the problem?????   Please give me some ideas.
    
    	Here are the accessing methods I tested.  (".." is double click)
    
    	Network Neighborhood
    	"Entire Network", Win95, Rawhide, Sable, Digital Cluster  (icons)
    	.. Digital Cluster
    	   .. AlphaCluster
    	      .. RawhideW, RawhideY, SableX, SableZ
    	         .. RawhideW .......................... Success!
    	         .. RawhideY .......................... Success!
    	         .. SableX   .......................... Failed!
    	            Access to the specified device, path, or file is denied.
    	         .. SableZ   .......................... Failed!
    	            Access to the specified device, path, or file is denied.
    
    	Network Neighborhood
    	"Entire Network", Win95, Rawhide, Sable, Digital Cluster
    	.. Sable
    	   .. SableX, SableZ
    	      .. SableX ............................... Success!
    	      .. SableZ ............................... Success!
    
    	This problem yields that everytime when I do the failover from 
    Rawhide to Sable, the Cluster Administrator software tells that the 
    failover was success.   I checked the disk icons in File Manager, it was 
    perfect, the drives were successfully failover.   Anyway, the client still 
    cannot see the shared resource anymore because it can't access the cluster 
    thru the Sable.
    	Any advice and recommendations would be greatly appreciated.
    
    Best regards,
    akkamon
    Technical Support - Windows NT
    
    Digital Thailand

    I've just figured out the problem was because of I mistyped the Cluster
    Alias name so that the cluster alias was the same as the domain name!
    
    regards,
    akkamon