| Author |
Message |
Dan
Guest
|
 Posted:
Thu Dec 16, 2004 10:33 am Post subject:
A clustering question |
|
|
Here is what I would like to do. I want to cluster (mirror?) 2 servers
together but have the servers located on different floors of the building so
that if one server goes down the other server automatically takes over
without any user intervention. I want this to be seamless to the client, in
other words, they will have no idea that one of the servers crashed.
Can I accomplish this by building 2 identical servers, same processor,
memory, RAID configuration, with a fiber link between the 2 for the private
network and a copper link for the public network? The servers will be
supporting a SQL application.
I do not want to use 2 nodes with shared storage because I need to have
complete failover for all hardware avoiding any downtime.
Is this at all possible using Microsoft's clustering? |
|
| Back to top |
|
 |
Gerald Aigenbauer
Guest
|
| Posted:
Thu Dec 16, 2004 2:44 pm Post subject:
Re: A clustering question |
|
|
hi dan!
the sync of the data discs between the two servers canīt be done by the
cluster server. cluster 2003 can only "sync" the quorum resource (majority
node set). you have to sync the data-disks via your hardware solutions. but
if you can manage that sync, then it will work.
gerald aigenbauer.
"Dan" <dmueller@netspheretech.com> schrieb im Newsbeitrag
news:2BC8447F-5B09-48FC-822A-9C1C1DB18612@microsoft.com...
| Quote: | Here is what I would like to do. I want to cluster (mirror?) 2 servers
together but have the servers located on different floors of the building
so
that if one server goes down the other server automatically takes over
without any user intervention. I want this to be seamless to the client,
in
other words, they will have no idea that one of the servers crashed.
Can I accomplish this by building 2 identical servers, same processor,
memory, RAID configuration, with a fiber link between the 2 for the
private
network and a copper link for the public network? The servers will be
supporting a SQL application.
I do not want to use 2 nodes with shared storage because I need to have
complete failover for all hardware avoiding any downtime.
Is this at all possible using Microsoft's clustering?
|
|
|
| Back to top |
|
|
Rodney R. Fournier [MVP]
Guest
|
| Posted:
Thu Dec 16, 2004 7:07 pm Post subject:
Re: A clustering question |
|
|
You have several ways to do this with SQL Server. Log shipping, using a
publisher/subscriber, standby server, and third party software. To get more
details on these, check the SQL cluster newsgroup or the Books on Line.
Cheers,
Rod
MVP - Windows Server - Clustering
http://www.nw-america.com - Clustering
http://www.msmvps.com/clustering - Blog
"Dan" <dmueller@netspheretech.com> wrote in message
news:2BC8447F-5B09-48FC-822A-9C1C1DB18612@microsoft.com...
| Quote: | Here is what I would like to do. I want to cluster (mirror?) 2 servers
together but have the servers located on different floors of the building
so
that if one server goes down the other server automatically takes over
without any user intervention. I want this to be seamless to the client,
in
other words, they will have no idea that one of the servers crashed.
Can I accomplish this by building 2 identical servers, same processor,
memory, RAID configuration, with a fiber link between the 2 for the
private
network and a copper link for the public network? The servers will be
supporting a SQL application.
I do not want to use 2 nodes with shared storage because I need to have
complete failover for all hardware avoiding any downtime.
Is this at all possible using Microsoft's clustering?
|
|
|
| Back to top |
|
|
kk
Guest
|
| Posted:
Tue Jan 04, 2005 6:48 am Post subject:
Re: A clustering question |
|
|
If you wait for Yukon you will also be able to use mirroring and
completely synchronize the SQL application with synchronous replication.
This will virtually mirror the SQL application on the 2 servers.
If you can't wait I would recommend a clustering solution like VERITAS
Cluster server along with Volume Replicator which can create a
replicated data cluster (RDC) which can perform the exact task that you
need for any clusterable application.
-kk
Rodney R. Fournier [MVP] wrote:
| Quote: | You have several ways to do this with SQL Server. Log shipping, using a
publisher/subscriber, standby server, and third party software. To get more
details on these, check the SQL cluster newsgroup or the Books on Line.
Cheers,
Rod
MVP - Windows Server - Clustering
http://www.nw-america.com - Clustering
http://www.msmvps.com/clustering - Blog
"Dan" <dmueller@netspheretech.com> wrote in message
news:2BC8447F-5B09-48FC-822A-9C1C1DB18612@microsoft.com...
Here is what I would like to do. I want to cluster (mirror?) 2 servers
together but have the servers located on different floors of the building
so
that if one server goes down the other server automatically takes over
without any user intervention. I want this to be seamless to the client,
in
other words, they will have no idea that one of the servers crashed.
Can I accomplish this by building 2 identical servers, same processor,
memory, RAID configuration, with a fiber link between the 2 for the
private
network and a copper link for the public network? The servers will be
supporting a SQL application.
I do not want to use 2 nodes with shared storage because I need to have
complete failover for all hardware avoiding any downtime.
Is this at all possible using Microsoft's clustering?
|
|
|
| Back to top |
|
|
|
|
|
|
FAQ
Memberlist
Register
Profile
Log in to check your private messages
Log in