Cluster setup question

#1

I'm getting ready to set up my first SQL Server 2005 Cluster so if I
sound like I don't really know what I'm doing...it's because I don't.
I'm reading white papers and doing research.

My plan is to have a cluster containing 2 nodes that are connected to
a shared disk array (DAS) with Raid 10 and a Raid 1 set for Logs and a
Raid 1 set for tempdb. The servers will have identical specs. My
plan was to create 3 instances (Production, TEST and DEV) and have the
server failover from the primary node to the secondary node in the
event of a hardware failure. This scenario would have the secondary
node basically doing nothing but waiting for its "moment in the sun"
should the primary node fail.

A question was raised about the posibility of having Production run on
the cluster and then having TEST and DEV run on a non-clustered
instance on Server B. This way Server B would be utilitized and
Server A would be spared the TEST and DEV workload at the processor
and ram level. There would be no failover for TEST and DEV in the
event Server B went down but they're not really mission critical. Is
this a terrible idea? Is such a setup even possible or should the
cluster be viewed exclusively as a single virtual server? It's not
clear to me if non-clustered instances can exist on a server that is a
node in a cluster.

Thanks in advance.

#2

It is possible to have clustered and non-clustered servers on the same host
node, but it is not advisable. The point of a cluster is to have high
availability. Given the complexity of a mixed environment, you may be
lowering the overall availability rather than increasing it.

--
Geoff N. Hiten
Senior SQL Infrastructure Consultant
Microsoft SQL Server MVP

"davisutt" <davisutt (AT) aol (DOT) com> wrote

Quote:

I'm getting ready to set up my first SQL Server 2005 Cluster so if I
sound like I don't really know what I'm doing...it's because I don't.
I'm reading white papers and doing research.

My plan is to have a cluster containing 2 nodes that are connected to
a shared disk array (DAS) with Raid 10 and a Raid 1 set for Logs and a
Raid 1 set for tempdb. The servers will have identical specs. My
plan was to create 3 instances (Production, TEST and DEV) and have the
server failover from the primary node to the secondary node in the
event of a hardware failure. This scenario would have the secondary
node basically doing nothing but waiting for its "moment in the sun"
should the primary node fail.

A question was raised about the posibility of having Production run on
the cluster and then having TEST and DEV run on a non-clustered
instance on Server B. This way Server B would be utilitized and
Server A would be spared the TEST and DEV workload at the processor
and ram level. There would be no failover for TEST and DEV in the
event Server B went down but they're not really mission critical. Is
this a terrible idea? Is such a setup even possible or should the
cluster be viewed exclusively as a single virtual server? It's not
clear to me if non-clustered instances can exist on a server that is a
node in a cluster.

Thanks in advance.

#3

It is possible to have clustered and non-clustered servers on the same host
node, but it is not advisable. The point of a cluster is to have high
availability. Given the complexity of a mixed environment, you may be
lowering the overall availability rather than increasing it.

--
Geoff N. Hiten
Senior SQL Infrastructure Consultant
Microsoft SQL Server MVP

"davisutt" <davisutt (AT) aol (DOT) com> wrote

Quote:

I'm getting ready to set up my first SQL Server 2005 Cluster so if I
sound like I don't really know what I'm doing...it's because I don't.
I'm reading white papers and doing research.

My plan is to have a cluster containing 2 nodes that are connected to
a shared disk array (DAS) with Raid 10 and a Raid 1 set for Logs and a
Raid 1 set for tempdb. The servers will have identical specs. My
plan was to create 3 instances (Production, TEST and DEV) and have the
server failover from the primary node to the secondary node in the
event of a hardware failure. This scenario would have the secondary
node basically doing nothing but waiting for its "moment in the sun"
should the primary node fail.

A question was raised about the posibility of having Production run on
the cluster and then having TEST and DEV run on a non-clustered
instance on Server B. This way Server B would be utilitized and
Server A would be spared the TEST and DEV workload at the processor
and ram level. There would be no failover for TEST and DEV in the
event Server B went down but they're not really mission critical. Is
this a terrible idea? Is such a setup even possible or should the
cluster be viewed exclusively as a single virtual server? It's not
clear to me if non-clustered instances can exist on a server that is a
node in a cluster.

Thanks in advance.

#4

It is possible to have clustered and non-clustered servers on the same host
node, but it is not advisable. The point of a cluster is to have high
availability. Given the complexity of a mixed environment, you may be
lowering the overall availability rather than increasing it.

--
Geoff N. Hiten
Senior SQL Infrastructure Consultant
Microsoft SQL Server MVP

"davisutt" <davisutt (AT) aol (DOT) com> wrote

Quote:

I'm getting ready to set up my first SQL Server 2005 Cluster so if I
sound like I don't really know what I'm doing...it's because I don't.
I'm reading white papers and doing research.

My plan is to have a cluster containing 2 nodes that are connected to
a shared disk array (DAS) with Raid 10 and a Raid 1 set for Logs and a
Raid 1 set for tempdb. The servers will have identical specs. My
plan was to create 3 instances (Production, TEST and DEV) and have the
server failover from the primary node to the secondary node in the
event of a hardware failure. This scenario would have the secondary
node basically doing nothing but waiting for its "moment in the sun"
should the primary node fail.

A question was raised about the posibility of having Production run on
the cluster and then having TEST and DEV run on a non-clustered
instance on Server B. This way Server B would be utilitized and
Server A would be spared the TEST and DEV workload at the processor
and ram level. There would be no failover for TEST and DEV in the
event Server B went down but they're not really mission critical. Is
this a terrible idea? Is such a setup even possible or should the
cluster be viewed exclusively as a single virtual server? It's not
clear to me if non-clustered instances can exist on a server that is a
node in a cluster.

Thanks in advance.

#5

It is possible to have clustered and non-clustered servers on the same host
node, but it is not advisable. The point of a cluster is to have high
availability. Given the complexity of a mixed environment, you may be
lowering the overall availability rather than increasing it.

--
Geoff N. Hiten
Senior SQL Infrastructure Consultant
Microsoft SQL Server MVP

"davisutt" <davisutt (AT) aol (DOT) com> wrote

Quote:

I'm getting ready to set up my first SQL Server 2005 Cluster so if I
sound like I don't really know what I'm doing...it's because I don't.
I'm reading white papers and doing research.

My plan is to have a cluster containing 2 nodes that are connected to
a shared disk array (DAS) with Raid 10 and a Raid 1 set for Logs and a
Raid 1 set for tempdb. The servers will have identical specs. My
plan was to create 3 instances (Production, TEST and DEV) and have the
server failover from the primary node to the secondary node in the
event of a hardware failure. This scenario would have the secondary
node basically doing nothing but waiting for its "moment in the sun"
should the primary node fail.

A question was raised about the posibility of having Production run on
the cluster and then having TEST and DEV run on a non-clustered
instance on Server B. This way Server B would be utilitized and
Server A would be spared the TEST and DEV workload at the processor
and ram level. There would be no failover for TEST and DEV in the
event Server B went down but they're not really mission critical. Is
this a terrible idea? Is such a setup even possible or should the
cluster be viewed exclusively as a single virtual server? It's not
clear to me if non-clustered instances can exist on a server that is a
node in a cluster.

Thanks in advance.

#6

It is possible to have clustered and non-clustered servers on the same host
node, but it is not advisable. The point of a cluster is to have high
availability. Given the complexity of a mixed environment, you may be
lowering the overall availability rather than increasing it.

--
Geoff N. Hiten
Senior SQL Infrastructure Consultant
Microsoft SQL Server MVP

"davisutt" <davisutt (AT) aol (DOT) com> wrote

Quote:

I'm getting ready to set up my first SQL Server 2005 Cluster so if I
sound like I don't really know what I'm doing...it's because I don't.
I'm reading white papers and doing research.

My plan is to have a cluster containing 2 nodes that are connected to
a shared disk array (DAS) with Raid 10 and a Raid 1 set for Logs and a
Raid 1 set for tempdb. The servers will have identical specs. My
plan was to create 3 instances (Production, TEST and DEV) and have the
server failover from the primary node to the secondary node in the
event of a hardware failure. This scenario would have the secondary
node basically doing nothing but waiting for its "moment in the sun"
should the primary node fail.

A question was raised about the posibility of having Production run on
the cluster and then having TEST and DEV run on a non-clustered
instance on Server B. This way Server B would be utilitized and
Server A would be spared the TEST and DEV workload at the processor
and ram level. There would be no failover for TEST and DEV in the
event Server B went down but they're not really mission critical. Is
this a terrible idea? Is such a setup even possible or should the
cluster be viewed exclusively as a single virtual server? It's not
clear to me if non-clustered instances can exist on a server that is a
node in a cluster.

Thanks in advance.

#7

It is possible to have clustered and non-clustered servers on the same host
node, but it is not advisable. The point of a cluster is to have high
availability. Given the complexity of a mixed environment, you may be
lowering the overall availability rather than increasing it.

--
Geoff N. Hiten
Senior SQL Infrastructure Consultant
Microsoft SQL Server MVP

"davisutt" <davisutt (AT) aol (DOT) com> wrote

Quote:

I'm getting ready to set up my first SQL Server 2005 Cluster so if I
sound like I don't really know what I'm doing...it's because I don't.
I'm reading white papers and doing research.

My plan is to have a cluster containing 2 nodes that are connected to
a shared disk array (DAS) with Raid 10 and a Raid 1 set for Logs and a
Raid 1 set for tempdb. The servers will have identical specs. My
plan was to create 3 instances (Production, TEST and DEV) and have the
server failover from the primary node to the secondary node in the
event of a hardware failure. This scenario would have the secondary
node basically doing nothing but waiting for its "moment in the sun"
should the primary node fail.

A question was raised about the posibility of having Production run on
the cluster and then having TEST and DEV run on a non-clustered
instance on Server B. This way Server B would be utilitized and
Server A would be spared the TEST and DEV workload at the processor
and ram level. There would be no failover for TEST and DEV in the
event Server B went down but they're not really mission critical. Is
this a terrible idea? Is such a setup even possible or should the
cluster be viewed exclusively as a single virtual server? It's not
clear to me if non-clustered instances can exist on a server that is a
node in a cluster.

Thanks in advance.

#8

It is possible to have clustered and non-clustered servers on the same host
node, but it is not advisable. The point of a cluster is to have high
availability. Given the complexity of a mixed environment, you may be
lowering the overall availability rather than increasing it.

--
Geoff N. Hiten
Senior SQL Infrastructure Consultant
Microsoft SQL Server MVP

"davisutt" <davisutt (AT) aol (DOT) com> wrote

Quote:

I'm getting ready to set up my first SQL Server 2005 Cluster so if I
sound like I don't really know what I'm doing...it's because I don't.
I'm reading white papers and doing research.

My plan is to have a cluster containing 2 nodes that are connected to
a shared disk array (DAS) with Raid 10 and a Raid 1 set for Logs and a
Raid 1 set for tempdb. The servers will have identical specs. My
plan was to create 3 instances (Production, TEST and DEV) and have the
server failover from the primary node to the secondary node in the
event of a hardware failure. This scenario would have the secondary
node basically doing nothing but waiting for its "moment in the sun"
should the primary node fail.

A question was raised about the posibility of having Production run on
the cluster and then having TEST and DEV run on a non-clustered
instance on Server B. This way Server B would be utilitized and
Server A would be spared the TEST and DEV workload at the processor
and ram level. There would be no failover for TEST and DEV in the
event Server B went down but they're not really mission critical. Is
this a terrible idea? Is such a setup even possible or should the
cluster be viewed exclusively as a single virtual server? It's not
clear to me if non-clustered instances can exist on a server that is a
node in a cluster.

Thanks in advance.

#9

It is possible to have clustered and non-clustered servers on the same host
node, but it is not advisable. The point of a cluster is to have high
availability. Given the complexity of a mixed environment, you may be
lowering the overall availability rather than increasing it.

--
Geoff N. Hiten
Senior SQL Infrastructure Consultant
Microsoft SQL Server MVP

"davisutt" <davisutt (AT) aol (DOT) com> wrote

Quote:

I'm getting ready to set up my first SQL Server 2005 Cluster so if I
sound like I don't really know what I'm doing...it's because I don't.
I'm reading white papers and doing research.

My plan is to have a cluster containing 2 nodes that are connected to
a shared disk array (DAS) with Raid 10 and a Raid 1 set for Logs and a
Raid 1 set for tempdb. The servers will have identical specs. My
plan was to create 3 instances (Production, TEST and DEV) and have the
server failover from the primary node to the secondary node in the
event of a hardware failure. This scenario would have the secondary
node basically doing nothing but waiting for its "moment in the sun"
should the primary node fail.

A question was raised about the posibility of having Production run on
the cluster and then having TEST and DEV run on a non-clustered
instance on Server B. This way Server B would be utilitized and
Server A would be spared the TEST and DEV workload at the processor
and ram level. There would be no failover for TEST and DEV in the
event Server B went down but they're not really mission critical. Is
this a terrible idea? Is such a setup even possible or should the
cluster be viewed exclusively as a single virtual server? It's not
clear to me if non-clustered instances can exist on a server that is a
node in a cluster.

Thanks in advance.

#10

It is possible to have clustered and non-clustered servers on the same host
node, but it is not advisable. The point of a cluster is to have high
availability. Given the complexity of a mixed environment, you may be
lowering the overall availability rather than increasing it.

--
Geoff N. Hiten
Senior SQL Infrastructure Consultant
Microsoft SQL Server MVP

"davisutt" <davisutt (AT) aol (DOT) com> wrote

Quote:

I'm getting ready to set up my first SQL Server 2005 Cluster so if I
sound like I don't really know what I'm doing...it's because I don't.
I'm reading white papers and doing research.

My plan is to have a cluster containing 2 nodes that are connected to
a shared disk array (DAS) with Raid 10 and a Raid 1 set for Logs and a
Raid 1 set for tempdb. The servers will have identical specs. My
plan was to create 3 instances (Production, TEST and DEV) and have the
server failover from the primary node to the secondary node in the
event of a hardware failure. This scenario would have the secondary
node basically doing nothing but waiting for its "moment in the sun"
should the primary node fail.

A question was raised about the posibility of having Production run on
the cluster and then having TEST and DEV run on a non-clustered
instance on Server B. This way Server B would be utilitized and
Server A would be spared the TEST and DEV workload at the processor
and ram level. There would be no failover for TEST and DEV in the
event Server B went down but they're not really mission critical. Is
this a terrible idea? Is such a setup even possible or should the
cluster be viewed exclusively as a single virtual server? It's not
clear to me if non-clustered instances can exist on a server that is a
node in a cluster.

Thanks in advance.