I was reading the Microsoft white paper
(http://www.microsoft.com/technet/prodtechnol/sql/2000/maintain/failclus.mspx) on clustering and noticed something I didn't expect.

It seems to say that only one node of a cluster may be used at a time and
that it is not for (any kind of) load balancing. The only apparent advantage
over log-shipping being an automatic failover.

Is this correct? If so, it seems rather strange to me (yes I understand the

that is correct

A instance (= virtual server = a cluster group) is only online on 1 node in
a server cluster

You can however run a second instance (= a second virtual server = a second
cluster group) on another node, however this instance has its own disk
resources and its own database(s).

this is the "shared nothing" model from MSCS

the real limitation is that you cannot have two Windows Servers accessing
the same NTFS volume. This is a NTFS limitation. If you have more than one
Windows Server accessing the same NTFS volume, it will lead to NTFS
corruption and more than likely data-loss. The cluster service manages the
disk resources, so only one node in the cluster will have ownership, and
therefore access, to the NTFS volume.

I hope this explains it
Rgds,
Edwin.


[quoted text, click to view]
(http://www.microsoft.com/technet/prodtechnol/sql/2000/maintain/failclus.msp
x) on clustering and noticed something I didn't expect.
[quoted text, click to view]

If you are interested in some sort of multi-server load-balancing, check out
Federated Servers in Books Online.


http://msdn2.microsoft.com/en-us/library/aa174502(sql.80).aspx

http://msdn2.microsoft.com/en-us/library/ms190381.aspx

http://msdn.microsoft.com/msdnmag/issues/06/04/ClusterComputing/default.aspx


Anthony Thomas


--

[quoted text, click to view]
(http://www.microsoft.com/technet/prodtechnol/sql/2000/maintain/failclus.msp
x) on clustering and noticed something I didn't expect.
[quoted text, click to view]