Conference decwet::winnt-clusters

633.0. "Adding more disks" by MPOS01::naiad.mpo.dec.com::mpos01::cerling (I'[email protected]) Tue Feb 18 1997 09:42

	I am having some difficulty in adding a new disk to the cluster.  I am
	running Beta 1.1 on a 1000/4100 cluster that has a BA box.  This is a
	demo system.  It was initially installed with a single drive in the
	BA box.  I found another disk and decided to add it so that I would 
	have a little more flexibility in the demo.  I had done this with V1.0
	so I didn't think I would have any problem here.  I guess I was
	mistaken.

	Everything works fine on the disk that was there when the cluster
	was installed.  I have tried several things.

	1. Simply slide the disk in while the system is running and run
	   Disk Administrator.  This works fine on my workstation running
	   NT Server V4.0.  On my workstation, I get the message stating
	   that it has discovered a different configuration, I click OK,
	   and have access to the disk.  I get the message on the cluster,
	   but when I click OK, I get an 'empty' representation of the
	   disk, with the message 'configuration information not
	   available' in the 'empty' representation.  The representation
	   for the original disk is correct.

	2. Turn off the second node in the cluster.  Hardware reboot first
	   node.  Same symptoms.

	3. Turn off clustering.  Set cluster services to manual start.  Hard
	   reboot.  Now both disks come up with 'configuration information
	   not available'.

	4. Even tried defining an alias for it within the cluster software
	   when I had it running.  The cluster software recognized that the
	   disk was there; gave me the signature, even.  But even then it
	   was never fully recognized by the system.

	In all cases, from AlphaBIOS, I am able to see the disks fine on the
	shared KZP controller.  In fact, when I tried it with the backup
	node still running, the original shared disk showed up 'off-line',
	just like it should, since the backup had control of it.  Obviously,
	it is being recognized by hardware, and NT sees that it is there, but
	I can't get it to be fully recognized.

	I didn't see anything in the documentation, but it wouldn't be the
	first time I missed something.  Or, did I find a real bug?

tgc

    You see a new grayed-out box after you run Disk Administrator?  That's
    good.  Run Disk Administrator on the other server in the cluster if
    it's up.  Now you should have a new gray box on each Disk Administrator
    display.
    
    That means that NT has recognized the disk, but the cluster software
    prevents access to the disk until you tell the cluster software to put
    the disk "online" on one of the systems.
    
    Run the Cluster Administrator program.  Modify one of your cluster
    failover groups and put the new disk into the group.  That will bring
    the disk online on the system that is currently serving the failover
    group.  Now when you run Disk Administrator on that system, you should
    see the disk as normal.

	That was it!  I guess I don't remember doing those steps from V1.0.
	Of course, my memory is getting so bad I think I could hide my own
	Easter eggs!

	Thanks,

tgc