Hello,

I currently have a website which is based upon a SQL2k cluster. This is fine
for fail-over support and does what is required. However, we have no real
maintenance window so when it comes to service packing or carrying out any
other maintenance works (on the disk array etc) then we suffer down time.
What I'd like to do is to set up a warm standby system to that when needed
we can point the web boxes at a different SQL box and then carry out the
maintenance.
So, is this possible/sensible.
Server A is the cluster and B is the standby box. Set up log shipping from A
to B with minimal latency. Then when needing to carry out mainenance on A we
stop A and restore the final log to B. Then we bring B online as the new web
database. When the maintenance of A is finished we log ship from B to A and
literally reverse the process. This could probably be set up with scripts so
shouldn't involve huge amounts of work.
It does sound a little cumbersome but I'm looking to reduce don-time as much
as possible. Are there better alternatives? How do others manage SQL
redundancy for the smallest amount of down time?

Thanks.
J

May be i'm not getting your probelm right.

You can always install patches and maintenance
using your current configuration unless
you want to do maintenance on clustered disks.
Service packs should never be applied on clsutered
disks anyways.

Log shipping will not solve this issue.