Yup I set the quorum up on the shared device.
Seems to failover ok :)

One other thing I've noticed, the IP address I assigned for the cluster, does this change to the NIC of the node that has control of the cluster?
I can see its been assigned to the Node1 at the moment (which has control) I can't take it down to see if Node2 gets assigned the address as I'm doing a database restore from tape at the moment..

I'm keeping my fingers crossed that the IP address changes to whichever node is master (It's all going just TOO well so far :D )!!

Cheers

[Edited by mega_stream - 7/2/2002 6:27:54 PM]

Quick story, Raid5 array shafted gone forever.

The cluster services now won't start as the array's gone. I need to rebuild the array with new discs and recreate the cluster object and get it seen by both nodes.
My plan is to uninstall cluster services, then reinstall with a fresh array. My question is, to reinstall is it just add/remove windows components? I think it is but can't remember!

Cheers

John

it is...

Cant you just remove this disks under cluster admin?

David

Well, I probably COULD have done, but its in such a bu99ered mess I can't even get into Cluster admin!

Oh well, off to battle I go :D

to be honest I would start again from scratch.... have you asked yourself why raid 5 failed... Im assuming it wasnt just one disk that died?

David

Right, fitted 5 new drives re-installed cluster services, service packed, all appears to be ok.
Node 1 if up first takes control of the shared drive
Node 2 takes over if Node 1 is powered down

Is it correct that the live node has exclusive access to the shared drive, ie the 2nd node can't browse the shared drive when until the 1st node goes down.

Seems to work, I assume its right!

That is correct i think... Ill check tomorrow..

both access quorum disk?

David

Mega

Yup you are right. (Unless you have shares located on the cluster disk as you may be able to connect this way.)

As David said make sure both can see the quorum on the shared disk. We used to have to use specific drivers on the Dell servers we used to run it on (just worth checking) although on W2K you should be ok.

I thought that there was a cluster IP for the cluster Netbios Name.. and then an IP for each node in the cluster?

and seperate IP's for hearbeat cards (depending how its configured.. as we use x - over cables..)

Ill check on ours tomorrow as its bugging me, because I should know this... ours may be different as the main lan cards are both in fault tollerant (sp) teams.

David

right..

our cluster is going to be called fred, and has two nodes called err... 1 and 2

fred is 10.0.0.50 resolves to fred
1 is 10.0.0.51 resolves to 1
2 is 10.0.0.52 resolves to 2

so I presume it will continue to work so long as you are using 10.0.0.50

David