Tuesday, March 27, 2012

Cluster setup question

Im viewing the Cluster setup instructions at:
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/howtosql/ht
_clustering_51rm.asp
But it doesnt say where to install SQL. It seems to me that if I wanted to
have an Active/ Passive Cluster I would need to intsall it on both boxes? Is
this correct? Are there better links I should be using?
--
SQL2K SP3
TIA, ChrisRYou begin the installation from the node that currently owns the disk
resource you want SQL data files installed to. You can add additional disk
resources after the installation. The SQL installer adds code automagically
to the nodes you specify, provided they actually exist and are running at
install time. You can add or remove nodes from the SQL installation later
if you need to. Once SQL is installed, there is no difference between
installed nodes except for which one is actually running SQL at a given
moment.
--
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
Careerbuilder.com
I support the Professional Association for SQL Server
www.sqlpass.org
"ChrisR" <bla@.noemail.com> wrote in message
news:ugjtmqZ$EHA.2136@.TK2MSFTNGP10.phx.gbl...
> Im viewing the Cluster setup instructions at:
>
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/howtosql/ht
> _clustering_51rm.asp
> But it doesnt say where to install SQL. It seems to me that if I wanted to
> have an Active/ Passive Cluster I would need to intsall it on both boxes?
Is
> this correct? Are there better links I should be using?
> --
> SQL2K SP3
> TIA, ChrisR
>|||Well, that is close, but not quite accurate.
Even on a stand-alone installation, when you run the setup dialogue in
interactive mode, all that really happens is that a custom .iss file is
generated from your responses to the questions. Once all of the information
is gathered, setup then executes the setupsql with the newly created
setup.iss file as a parameter.
A cluster install is similar, except the setup.iss is deployed to both
nodes. During the installation, the cluster fails the disk over to each
node so that the individual setup.iss files can be ran against two
independently executed setupsql commands. This creates registry keys.
However, since the first install already created the directory structure,
deployed the binaries, installed the .sql scripts to master, model, and
msdb, this part of the process does not run again a second time. Just the
registry keys, performance counters, and client side tools. And, of course,
the MDAC, which is client side and IS NOT cluster aware.
Also, there is such a thing as the "Primary Node." This is nothing more
than the node from which you started the installation. It does not matter
which node you begin the installation on, as long as it is the node that
currently owns the disk resource you plan on installing SQL Server to.
However, it is important that you keep in mind which node this was.
This node will be the only one that maintains a record of the setup logs.
There is also special CLSIDs that only this node will contain that will not
propagate to the other. The nodes are close to the same configuration, BUT
NOT EXACT. This will be important in the future when you try to add
additional disk resources or swap out old disks for new, if you ever need to
modify or add an additional IP resource, or change the domain the cluster
belongs to. Keeping in mind which node was PRIMARY will only be to your
benefit.
Sincerely,
Anthony Thomas
"Geoff N. Hiten" <SRDBA@.Careerbuilder.com> wrote in message
news:%23e8GA4Z$EHA.2196@.TK2MSFTNGP14.phx.gbl...
You begin the installation from the node that currently owns the disk
resource you want SQL data files installed to. You can add additional disk
resources after the installation. The SQL installer adds code automagically
to the nodes you specify, provided they actually exist and are running at
install time. You can add or remove nodes from the SQL installation later
if you need to. Once SQL is installed, there is no difference between
installed nodes except for which one is actually running SQL at a given
moment.
--
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
Careerbuilder.com
I support the Professional Association for SQL Server
www.sqlpass.org
"ChrisR" <bla@.noemail.com> wrote in message
news:ugjtmqZ$EHA.2136@.TK2MSFTNGP10.phx.gbl...
> Im viewing the Cluster setup instructions at:
>
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/howtosql/ht
> _clustering_51rm.asp
> But it doesnt say where to install SQL. It seems to me that if I wanted to
> have an Active/ Passive Cluster I would need to intsall it on both boxes?
Is
> this correct? Are there better links I should be using?
> --
> SQL2K SP3
> TIA, ChrisR
>|||I have had to change accounts and add/remove disk resources. As long as I
use the current owner node, everything works fine.
Yes, there are some subtle differences such as log files between nodes but
that is only relevant during installation failure troubleshooting. Once
everythign is running, there is no practical difference between nodes,
unlike SQL 7.0 clustering.
Geoff N. Hiten
Microsoft SQL Server MVP
Senior Database Administrator
Careerbuilder.com
I support the Professional Association for SQL Server
www.sqlpass.org
"Anthony Thomas" <ALThomas@.kc.rr.com> wrote in message
news:OfrkU8e$EHA.1084@.tk2msftngp13.phx.gbl...
> Well, that is close, but not quite accurate.
> Even on a stand-alone installation, when you run the setup dialogue in
> interactive mode, all that really happens is that a custom .iss file is
> generated from your responses to the questions. Once all of the
information
> is gathered, setup then executes the setupsql with the newly created
> setup.iss file as a parameter.
> A cluster install is similar, except the setup.iss is deployed to both
> nodes. During the installation, the cluster fails the disk over to each
> node so that the individual setup.iss files can be ran against two
> independently executed setupsql commands. This creates registry keys.
> However, since the first install already created the directory structure,
> deployed the binaries, installed the .sql scripts to master, model, and
> msdb, this part of the process does not run again a second time. Just the
> registry keys, performance counters, and client side tools. And, of
course,
> the MDAC, which is client side and IS NOT cluster aware.
> Also, there is such a thing as the "Primary Node." This is nothing more
> than the node from which you started the installation. It does not matter
> which node you begin the installation on, as long as it is the node that
> currently owns the disk resource you plan on installing SQL Server to.
> However, it is important that you keep in mind which node this was.
> This node will be the only one that maintains a record of the setup logs.
> There is also special CLSIDs that only this node will contain that will
not
> propagate to the other. The nodes are close to the same configuration,
BUT
> NOT EXACT. This will be important in the future when you try to add
> additional disk resources or swap out old disks for new, if you ever need
to
> modify or add an additional IP resource, or change the domain the cluster
> belongs to. Keeping in mind which node was PRIMARY will only be to your
> benefit.
> Sincerely,
>
> Anthony Thomas
>
> --
> "Geoff N. Hiten" <SRDBA@.Careerbuilder.com> wrote in message
> news:%23e8GA4Z$EHA.2196@.TK2MSFTNGP14.phx.gbl...
> You begin the installation from the node that currently owns the disk
> resource you want SQL data files installed to. You can add additional
disk
> resources after the installation. The SQL installer adds code
automagically
> to the nodes you specify, provided they actually exist and are running at
> install time. You can add or remove nodes from the SQL installation later
> if you need to. Once SQL is installed, there is no difference between
> installed nodes except for which one is actually running SQL at a given
> moment.
> --
> Geoff N. Hiten
> Microsoft SQL Server MVP
> Senior Database Administrator
> Careerbuilder.com
> I support the Professional Association for SQL Server
> www.sqlpass.org
> "ChrisR" <bla@.noemail.com> wrote in message
> news:ugjtmqZ$EHA.2136@.TK2MSFTNGP10.phx.gbl...
> > Im viewing the Cluster setup instructions at:
> >
> >
>
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/howtosql/ht
> > _clustering_51rm.asp
> >
> > But it doesnt say where to install SQL. It seems to me that if I wanted
to
> > have an Active/ Passive Cluster I would need to intsall it on both
boxes?
> Is
> > this correct? Are there better links I should be using?
> >
> > --
> > SQL2K SP3
> >
> > TIA, ChrisR
> >
> >
>sqlsql

No comments:

Post a Comment