Actiuve/Active

Hi,
I have to setup a Active/Active cluster and I need to know if there has to
be made any change in the applications that the user has? For example, direct
the app to the other virtual server, etc...
Thanks
What do you mean by "Active-Active"? It sounds like you may be
misunderstanding what the terms mean. "Active-Active" describes an onder
clustering technology that went away after SQL 7.0. The current correct
term is "multi-instance" and acts just like having multiple instances of SQL
Server on a single, non-clustered host.
Geoff N. Hiten
Senior Database Administrator
Microsoft SQL Server MVP
"Fred" <Fred@.discussions.microsoft.com> wrote in message
news:A2B00583-DDC2-41A1-94D9-394F2E985C6B@.microsoft.com...
> Hi,
> I have to setup a Active/Active cluster and I need to know if there has to
> be made any change in the applications that the user has? For example,
> direct
> the app to the other virtual server, etc...
>
> Thanks
|||Sorry,
what I meant was, we're going in a process of installing a cluster using
something similar to what then was called "Active/Active" in a sense that one
server serves a set of databases a acts as a failover through a instance what
will be activated in case of a failure in the other server. Something like:
Server1: Actives databases: A, B, C
Inactives databases: D, E, F
Server2: Actives databases: D, E, F
Inactives databases: A, B, C
In case of a failure in Server1, the Server2, serves not only databases D,
E, F, but also A, B, C. I think that's called in the old terminology
"Active/Active". My major point is, do I have to make some sort of change in
my applications to "redirect" them to the "new server" or the cluster can
that care of that automatically?
Is there any source that really clarifies clustering technology in terms of
applications made by our programmers?
Thanks for your observations, point taken.
"Geoff N. Hiten" wrote:

> What do you mean by "Active-Active"? It sounds like you may be
> misunderstanding what the terms mean. "Active-Active" describes an onder
> clustering technology that went away after SQL 7.0. The current correct
> term is "multi-instance" and acts just like having multiple instances of SQL
> Server on a single, non-clustered host.
> --
> Geoff N. Hiten
> Senior Database Administrator
> Microsoft SQL Server MVP
>
>
> "Fred" <Fred@.discussions.microsoft.com> wrote in message
> news:A2B00583-DDC2-41A1-94D9-394F2E985C6B@.microsoft.com...
>
>
|||Ok, you do have an almost correct picture of SQL Clustering.
It works more like this:
Instance 1: Databases A, B, C
Instance 2: Databases D, E, F
Host machines: Node 1, Node 2
Normal operation has Instance 1 on Node 1 and Instance 2 on Node 2. If node
2 fails, Node 1 grabs the resources for Instance 2 and starts up Instance 2
on Node 1. The outside world still sees Instance 2 with the exact same
names, security, and IP address(es) as before. The two( or more depending
on your cluster configuration) instances look like completely different and
independent servers to the outside world. Reconnection is necessary since
the SQL service stopped and restarted, but no client redirection is required
since the clients connect to a virtual IP address.
Geoff N. Hiten
Senior Database Administrator
Microsoft SQL Server MVP
"Fred" <Fred@.discussions.microsoft.com> wrote in message
news:C20D7184-9C4B-4D63-B3F8-AA37951F3CC0@.microsoft.com...[vbcol=seagreen]
> Sorry,
> what I meant was, we're going in a process of installing a cluster using
> something similar to what then was called "Active/Active" in a sense that
> one
> server serves a set of databases a acts as a failover through a instance
> what
> will be activated in case of a failure in the other server. Something
> like:
> Server1: Actives databases: A, B, C
> Inactives databases: D, E, F
> Server2: Actives databases: D, E, F
> Inactives databases: A, B, C
> In case of a failure in Server1, the Server2, serves not only databases D,
> E, F, but also A, B, C. I think that's called in the old terminology
> "Active/Active". My major point is, do I have to make some sort of change
> in
> my applications to "redirect" them to the "new server" or the cluster can
> that care of that automatically?
> Is there any source that really clarifies clustering technology in terms
> of
> applications made by our programmers?
> Thanks for your observations, point taken.
>
> "Geoff N. Hiten" wrote:
|||Thanks, this post helped a lot to clarify these points to me.
"Geoff N. Hiten" wrote:

> Ok, you do have an almost correct picture of SQL Clustering.
> It works more like this:
> Instance 1: Databases A, B, C
> Instance 2: Databases D, E, F
> Host machines: Node 1, Node 2
> Normal operation has Instance 1 on Node 1 and Instance 2 on Node 2. If node
> 2 fails, Node 1 grabs the resources for Instance 2 and starts up Instance 2
> on Node 1. The outside world still sees Instance 2 with the exact same
> names, security, and IP address(es) as before. The two( or more depending
> on your cluster configuration) instances look like completely different and
> independent servers to the outside world. Reconnection is necessary since
> the SQL service stopped and restarted, but no client redirection is required
> since the clients connect to a virtual IP address.
>
> --
> Geoff N. Hiten
> Senior Database Administrator
> Microsoft SQL Server MVP
>
>
> "Fred" <Fred@.discussions.microsoft.com> wrote in message
> news:C20D7184-9C4B-4D63-B3F8-AA37951F3CC0@.microsoft.com...
>
>