It is a fact, that once we
setup an Active/Passive cluster and install SQL server on a cluster server, the
passive nodes will never be used unless it fails over. The main objective here
is to try to make use of the unused resource and equipment, without spoiling
the failover concept.
Objective:
The objective in simple words is
to make an Active/Passive SQL Cluster to a Quassi Active/Active SQL Cluster.
Environment:
Physical Servers SQLBOX1 [165.100.100.1], SQLBOX2
[165.100.100.2]
Virtual Servers VIRTUALBOX1[172.177.100.1],
VIRTUALBOX2[172.177.100.2]
VIRTUALBOX1 has SQLBOX1
as primary node and SQLBOX2 as secondary node.
Prerequisites:
We assume that an
active/passive cluster [VIRTUALBOX1] has already been setup between the
physical servers SQLBOX1 and SQLBOX2, and SQL Server has been running smoothly
on the VIRTUALBOX1 cluster.
Process:
- Get a new IP address for the
VIRTUALBOX2. -
Get new drives from the SAN for
SQLBOX2. -
Install SQL server on SQLBOX2 as the
primary node. Give the virtual IP address for the virtual server. -
SQL installation will ask for an
instance name. Call it SQL1. This will install SQL server instance
VIRTUALBOX2/SQL1 on that box.
SQL Server Access:
Now you can access the SQL
servers on VIRTUALBOX1 as VIRTUALBOX1 and VIRTUALBOX2 as VIRTUALBOX2/SQL1
Failover:
 Click for larger image |
When SQLBOX1 fails over to
SQLBOX2, there will be no changes in Virtualbox2, since SQLBOX2 is the active
node for it.
When SQLBOX2 fails over to
SQLBOX1 there will be no changes in VIRTUALBOX1 since SQLBOX1 is the active
node for it.
Hence by creating an Quassi Active/Active cluster server we will be using
all of the nodes and its resources. In addition, we will get two failover clustered sql server servers instead of one.
