RAC INTRODUCTION
Shared Storage
Box
INSTANCE
1 (SAN and NAS)
INST2 Database files
Or Or are
stored(DATABA
NODE1 Node2 SE)
Client s
Definition:
A group of individual systems connected that an acting as
logically one system is called clustering
In this RAC environment we must require mini two nodes having
individual instances and accessing common data from shared
storage box
Advantage:
1. High data availability <we can add more storage device as
shown above>
2. Zero down time <always servers are up and running>
3. Load balancing <distributing the load among the nodes>
UNDERSTANDING ABOUT SHARED STORAGE DEVICE:
This shared storage device contains data which can be accessed
by multiple nodes in RAC environment. This storage box can be in
the SAN environment or NAS.
SAN- storage area network NAS-network attached storage
SAN: In this environment …..This storage box accessed via the
SAN switch through the optical fiber cables
This optical fiber cables can transfer the data among the nodes
up to 32 GB
NAS: It can be accessed in the network by using hub or normal
switch through fire wired cables or catch5 cables<ordinary>
UNDERSTANDING ABOUT RAC ARCHITECTURE:
In RAC env, we require two nodes with two network adapters each
network adapter assign with ip addresses.
Note: public ip addresses are use full to make the connectivity
between the clients and nodes through public switch.
Private ip addresses must be assign to make the connection
between the storage box and nodes through the private switch.
C1-clients
C1 C2 C3
Public ip address
Node1:192.168.1.101
Public
NICcard sswitc Node2:192.168.1.102
h through public switch
Private Ip
Node Private Node address
switch 2
1 Node1:192.168.2.101
Node2:192.168.2.102
through private switch
NIC-Network Interface
San box it
contains
Database files
and ocr &voting
What is difference between stand alone database and RAC
files
data base:
STAND ALONE DATA BASE: (only one instance)
• Only one instance with set background processes
• Require minimum two redo log groups with at least one
redologfile in each group
• It also requires at least one undo table space
• It does not support high availability, high scalability and
load balancing
Note: scalability means we can add a node or delete a node
without disturbing another node to expand the network.
RAC database:
• Require mini two nodes
• Each node contains own SGA & PGA and set of background
processes
Note: apart from the normal background processes …every node
contains the following additional background processes.
LMON, LMSN, LOCK, DIAG and LMD.
• Mini two redo log groups for each instances
• It supports high scalability, high data availability, high
performance and load balancing.
Another concept is buffering the data between the two instances
for user requirements………for that oracle corporation uses oracle
cache fusion
Oracle cache fusion
• It is building soft with all RAC versions from 9i
• It is use full to synchronize the buffers from one
instance to another instance this process is called
buffer synchronization
• Using this high data availability is possible
• It also improves the database performance while
accessing the records from the buffers
CONCURRENCY, CONSISTENCY, COHERENCY:
CONCURENCY: Means that modifying the data is available in all
the nodes at any time by using this process we can achieve high
availability
CONSISTENCY: it is also similar to concurrency
But consistency means each instance must be synchronized
dynamically from other nodes in RAC environment to provide the
ZERO DOWN TIME
COHERENCY:
It is process to monitor consistence of data in all the nodes in
RAC environment
RAC BACKGROUND PROCESSES:
Every instance contains its own background processes and RAC
has extra Background processes they are LMON, LMSN, LOCK,
DIAG, and LMD.
LMON: global enque service monitor
This is useful to monitor the connectivity between the two nodes
and monitoring other background processes activities.
LMSN: global cache service processes
This is useful to identify the modified blocks in the GRD buffer
Grd-global resource directory
LOCK:
This is useful to keep the lock on update transactions till
transaction is completed
DIAG:
Diagnostic base background processes is useful to generate log
and trace files to monitor the database activities
LMD: global enque service Daemon
This is useful to monitor the all the cluster database activities
between the nodes
RAC supports load balancing
LOADBALANCING: whenever no of user requests are coming to
the nodes, at that time some of the userrequeats goes to one
node, other request goes to other node and user request will be
distributed among the nodes in RAC env is called load balancing.
RAC COMPONENTS:
GLOBAL CACHE SERVICE:
Is useful to identify the modified data blocks in the buffer
GLOBAL ENQUE SERVICE:
Is useful to apply logs on the modified data blocks with help of
lock processes till transaction is not completed
GLOBAL RESOURCE DIRECTORY<GRD>:
Every instance contains virtual global resource directory which is
useful to hold the modified blocks information
