Basic Clustering and Failovers

I'm thinking of setting up clustering and have a basic failover question that I can't seem to find and an answer.

I want to setup two sql 2005 std ed servers and have them clustered. When the servers are running are they both taking requests (load balanced) or is only one server taking requests and the other being updated, waiting for the first server to fail. If this is true, if/when the first server fails does the second one step up automatically?

I couldn't find any resources that answered these seemingly beginner questions. I'm more than happy to read so if you know of any articles/books I'd appreciate that as well.

thanks in advance,

Scott

Scott,

First, you have to install windows cluster server before starting to install sql server.

Running SQL Server setup detetcts then cluster and installs sql server on all nodes.

For SQL Server There are two possible setups

a) active/passive = both servers are running. one is the active (taking requests) the other is standby to takeover.

b) active/active = both nodes are taking request (each server for it's own db's) in case one brakes down the other takes all load.

A loadbalancing setup is not possible.

HTH

acki

|||

Thanks, oddly enough for some reason I just couldn't find info for how the srevers functioned once it was installed.

Quick follow-up--for option b (Active/Active), how do you access the server from the name of the cluster and/or the individual server names?

|||When clustering SQL Servers you always access the SQL Server via the virtual name. In an Active/Active config each SQL Instance that you install has a unique hostname and a unique instance name (one can be default and one must be a named instance, or both can be named instances, but they must be different names).

Backups

Hi All,
We have setup Transactional Replication with queued updating subscribers option. Our Servers are in Failover cluster with windows 2000 AS. We are doing Full and Transactional log backups as usual. Can any one suggest what are the recovery scenarios, in ca
se of Publisher (if it is completely down, I mean all clusters are down) and User interactions to Subscriber immediately, so that data is more than of Publisher db.
Please provide us valuable suggestions.
Thanks,
John.
I'm not exactly sure what your question is. You loose me with this statement
:
"so that data is more than of Publisher db. "
With queued Updating, your Subscriber is essentially a mirror of your
Publisher's data. If your Publisher goes offline permanently you will lose
transactions that are on your publisher and haven't made their way back to
the Subscriber.
So, its not too hard to clone your Subscriber to rebuild your Publisher. To
do this you would have to drop your subscriber, copy the subscriber schema
and data to the publisher, re run your publication scripts on your new
publisher, and do a no-sync subscription.
If your Publisher goes offline for an extended period your chance of
bringing it back on line without conflicts increases. With Queued Updating
the conflict view only lets you view conflicts, not resolve them.
"John" <anonymous@.discussions.microsoft.com> wrote in message
news:1D2A989A-5347-4E74-AAD6-707E9EC9CE70@.microsoft.com...
> Hi All,
> We have setup Transactional Replication with queued updating subscribers
option. Our Servers are in Failover cluster with windows 2000 AS. We are
doing Full and Transactional log backups as usual. Can any one suggest what
are the recovery scenarios, in case of Publisher (if it is completely down,
I mean all clusters are down) and User interactions to Subscriber
immediately, so that data is more than of Publisher db.
> Please provide us valuable suggestions.
> Thanks,
> John.
>
|||The best suggestion I have is to look in BOL for the article about the
"synch with backup" option that can be used. This will not only walk you
through these scenarios, but explain how to create coherent, multi-server
backups.
Mike
Principal Mentor
Solid Quality Learning
"More than just Training"
SQL Server MVP
http://www.solidqualitylearning.com
http://www.mssqlserver.com