You are on page 1of 4

Basic cluster look:

SAN

N1 PN N2

Configuration:
Node1 node2
Private network 10.10.10.1 10.10.10.2
Public network 200.200.200.1/8 200.200.200.2/24
8=subnet mask 24=subnet mask
255.0.0.0 255.255.255.0

Windows cluster name : Winclus (name to address the entire cluster)


Windows cluster IP 200.200.200.10

Disk group 1:
Q:\
Cluster name
Cluster IP

Disk group 2:
X:\

Points:
1. Even if windows resources is owned by node 1, SQL installation can be done
from other node node 2
But the SQL resource i.e. Disk for SQL installation should be owned by node 2.
2. For named instance we give a new name to the SQL Server cluster instance
instead of giving the SQL Network name itself because there will be a confusion
For example: if SQL Network Name (SQL Virtual Server name) is Assembly1
And if we give Instance name also as Assembly1 then we will
reference the instance by Assembly1\Assembly1 which might create
confusion to customer as well as to PSS Engineer

3. SQL cluster name cannot be same as windows cluster name because both will
have different IP address and it will be difficult to map the two IP address.

4. to identify whether the installation was a default or named instance,

For default for named


Network name (SQLCVN) network name (SQLCVN)
IP Address (SQLCVN) IP Address (SQLCVN)
SQL Server SQL Server (instance1)
SQL Agent SQL Agent (instance 1)
Full text search Full text search (instance 1)

If you see some name (instance1) in SQL Server resource then it’s a named
instance. No name in SQL Server resource indicates default instance

5. If we have a SQL cluster with an instance as SQLVCN\instance1, to install next


instance of SQL Server you cannot use same virtual server name as SQLVCN.
Every installation of SQL Server needs a new SQL virtual server name. It is just
like giving a name to a pair of IP address on a cluster.

There is a mapping between SQL Server Virtual Server name and IP


So every time we install a new instance we need to give a new and unique Virtual
Server Name since every new instance will have a unique IP

6. We cannot use the same instance name even if we give different SQL Network
Name (SQL Server Virtual Server name) because the binaries will be created in
MSSQL\instance 1 for the server SQLVCN1\instance 1 and also at the same
location for SQLVCN2\instance 1 on either of the nodes. This is not possible.
Example: SQLVCN1\instance1 (First instance)
SQLVCN1\instance2 – Not correct since SQL Network name cannot be
same
SQLVCN2\instance1 – Not correct since SQL instance name cannot be
same
SQLVCN2\instance2 – Correct since SQL network name and instance
name are different and unique
7. A cluster installation is not possible in a workgroup since we need a user who can
be made a local admin on both the machines which can be a domain user account
because user of node1 is invisible for node2.

8. Changing of SQL network name is not possible at all.

9. Changing of SQL IP address is possible by running the set up again and if you
have already made changes in IP using the cluadmin, then go and make changes
in the registry also.
HKLM\software\Microsoft\Microsoft SQL Server\cluster
HKLM\software\Microsoft\MSSQLServer\instancename\cluster

Steps for installation of a cluster:

1. Install Active Directory.


2. Install nodes.
3. On the nodes activate cluster services
4. Create a windows cluster.
5. Start installation of SQL Server.

Dependencies:

Agent services full text services

SQLServer

SQL server Disk


network name

SQL IP address

Meaning:
Signifies “depends on”

You might also like