|
Database and Database Instance
Oracle Database Server represents a collection of
physical files, logical database objects such as
tables and indexes, and the host level memory
structures and processes. The physical host level
files reside on storage arrays, directly attached,
network attached, or on a storage area network
(SAN). The combination of background processes and
memory buffers is called the database instance. An
Oracle database instance, which resides on a host,
is the actual database processing area that allows
access to the physical and logical structures.
A RAC database consists of multiple database
instances.
Access to the database (dbf files) is shared
by the multiple instances.
In other words, the database consists of a
single set of physical data files that can be
accessed by multiple database instances.
As shown in Figure 2.1, each of the instances
resides on a separate host and forms its own set of
background processes and memory buffers. Thus, RAC
enables access to a single database via multiple
database instances.
instance_name vs. db_name
When the database is not a RAC system, it has one
instance and one database. Sometimes, the instance
and database are construed to be the same. In that
case, it is called a standalone database system.
As an example,
Database Name
: NYDB50
Instance-1 Name
: NYDB51
Instance-2 Name
: NYDB52
Instance-3 Name
: NYDB53
The parameter
db_name
will have the value of NYDB50; this represents the
name of the database.
The parameter
instance_name
will be one of the names listed above. All of these
instances provide access to the same database named
NYDB50.
Figure 2.1:
Multi-Instance
RAC Database System – At a Glance
Database Instance
A typical RAC instance is similar to a stand-alone
instance.
A RAC instance has extra processes, memory
structures, and logical structures. Since RAC must
maintain concurrency of data across multiple
instances, it creates additional structures to
manage and coordinate the resources.
System Global Area (SGA)
The SGA components include the database buffer
cache, large pool, java pool, streams pool, redo log
buffer and the shared pool.
The main memory structures of the shared pool
are the dictionary cache, library cache and the
result cache.
Unlike the PGA, the SGA memory structures are
shared.
The SGA and PGA memory structures are shown in the
Figure 2.2. The SGA resources are formed at the time
of database instance launch based on the instance
initialization parameters. However, many of the
parameters can dynamically be altered to suit the
database processing needs.
Figure 2.2:
SGA and PGA
Components of a Typical Instance
|
