"Linchi Shea" <LinchiShea@discussions.microsoft.com> wrote in message
news:A5BB060D-037D-4CEF-BE98-4EBEC37A13AF@microsoft.com...
> Call me paranoid. But I'd suggest that you do plan for some down time for
> this upgrade just in case. I'm not saying that the rolling replacement
> doesn't work. I'm saying this from experience and observation. One
> observation is that even our most critical
> if-it's-down-the-company-is-out-of-business application can find
> maintenance
> time (such as annual power down etc) to piggyback its upgrade, and even
> though the vendor (not MS) supports rolling upgrade (not SQL Server) in
> tightly controlled hardware environment, people don't really want to take
> the
> risk of dealing with the complexity.
>
> Just being cautious here. If you really can't plan any down time, you've
> got
> to what you've got to do.
>
> Linchi
>
> "Replacing a cluster node" wrote:
>
>> Hello to all.
>>
>> We have a cluster server in active-pasive configuration, that is running
>> sql
>> server 2000. We are planning replace both nodes of the cluster (one at
>> each
>> step) without taking the entire cluster offline, and it´s critical that
>> the
>> sql service don´t stop when the replacement is being made.
>>
>> Our plan to replace each node is:
>>
>> 1) Move all the groups to the node that is not being replaced, shut down
>> the
>> node to be replaced, disconnect it, then connect the new node.
>> 2) Connect the new node and restore the cluster node configuration data
>> and
>> configuration into the new node using Automated System Recovery from the
>> backups previously made from the old node.
>> 3) Install sql server on the new node (only installing the database
>> server
>> and ataching it to the existing virtual server as described in
>> http://www.microsoft.com/technet/prodtechnol/sql/2000/maintain/failclus.mspx
>> in "Adding or Removing a Cluster Node from the Virtual Server
>> Definition").
>>
>> Our question is simple: Our plan will work? Any suggestions?
>>
>> Thanks
>>
>>