3 node RAC on Enterprise Linux
ocr, voting disk, spfile on raw devices
and the db files on the block device
storage on MSA1000 ( modular smart array) with 2 fc and gigabit ethernet card .
public eth0 , private eth1 , vip eth0:1
-set up user equivalence.
-checking the configuration of hangcheck timer module
-check the rawdevice mapping exists. for ocr,voting and asm spfile e.g raw -qa
-check that the ASM disk are configured for you -these are block device so you upgrade oracle software from 10.2.0.1 to 10.2.0.3 - (right? )
- install 10.2.0.1 clusterware
-patch oracle clusterware to 10.2.0.3 (rolling clusterware patch)
-install db software in one home to be used by both asm and db.
-patch to 10.2.0.3
-Create cluster database with dbca
- choose RAC database
- choose create database
- select all nodes in the cluster
- select general purpose db
- choose a db name prod
- uncheck configure the db for enterprise manager
- specify sys as password for all account
- choose "Automatic storage management(ASM)" as the storage mechanism
- Configuring ASM - since ASM has not yet been configured dbca allows you to do this at this point of database creation
-specify sys as the asm SYS password
-specify our ASM spfile raw device as the spfile for ASM.
-click the create new button to create a new ASM diskgroup DATA
-create another ASM diskgroup FRA
- choose oracle managed file and the +DATA diskgroup
- select speficy flashrecovery area and +FRA diskgroup
- enable archivelog if you want
- on the database services screen create 2 services
"oltp" preferred node1,node2 available node3
"batch" preferred node3 , available node1,node2
- continue clicking default till it finish .
and now " my second RAC install is finished "
Wednesday, April 30, 2008
my First RAC experience
This week was amazing for me . on monday I did a 2 node RAC installation on solaris and on Tuesday I did a 3 node RAC on enterprise Linux .
2 node RAC10g on Solaris 10.
two v240 server with EMC clarion storage.
some routine work - user equivalence check , soft link to /usr/local/bin for ssh and scp . make sure your storage lun is outlined. then the public bge0 ,private bge1 ,vip bge0:1 address. (vip is configured by vipca at the end of clusterware. ..)
1- clusterware installation
2- ASM installation on a asm home
2.1 - As part of asm installation on the select configuration option - choose configure Automatic Storage management(ASM)
that leads you to define diskgroup . selected the candidate raw partitions . and the rest follows
3- database installation on a separate db home
3.1 - on the select configuration option choose only the Install database software only .
4- Create the cluster database
4.1 - run dbca- choose Oracle Real Application Clusters database -
on the Select the operation that you want to perform page
choose Configure Automatic Storage Management
on the Node select choose select all
enter the ASM sys user password
on the ASM disk groups it will allow you to create another FRA diskgroup if you want
select the raw parttions that you want to assigned to FRA
you should see for all the diskgroups the state shows mounted(num nodes/num nodes )
Click Finish and it ask you "Do you want to perform another operation ?" Choose Yes
Back to the operations screen where you select database .
on the node selection choose select all...
defined a db template- General purpose
on the management options enter the database name and the sid prefix.
then on step 6 of 16 use the same password for all account
On the storage option screen . select Automatic storage management -
select the DATA diskgroup to be used for the db storage . uncheck FRA
On the database files location screen . select oracle-managed file
Select the specify flash recovery area check box and enter FRA diskgroup to be used for fla
select some sample schema if you please
Create database service if you like ( to differentiate based on usage , batch for 1 node,oltp for both
then click , click
on the Creation option page select the create database option .
then click click click and you got a window
Starting cluster database "DEVL" and its instances "DEVL1,DEVL2 " in progress. ...
and now < I am done with my First RAC >
2 node RAC10g on Solaris 10.
two v240 server with EMC clarion storage.
some routine work - user equivalence check , soft link to /usr/local/bin for ssh and scp . make sure your storage lun is outlined. then the public bge0 ,private bge1 ,vip bge0:1 address. (vip is configured by vipca at the end of clusterware. ..)
1- clusterware installation
2- ASM installation on a asm home
2.1 - As part of asm installation on the select configuration option - choose configure Automatic Storage management(ASM)
that leads you to define diskgroup . selected the candidate raw partitions . and the rest follows
3- database installation on a separate db home
3.1 - on the select configuration option choose only the Install database software only .
4- Create the cluster database
4.1 - run dbca- choose Oracle Real Application Clusters database -
on the Select the operation that you want to perform page
choose Configure Automatic Storage Management
on the Node select choose select all
enter the ASM sys user password
on the ASM disk groups it will allow you to create another FRA diskgroup if you want
select the raw parttions that you want to assigned to FRA
you should see for all the diskgroups the state shows mounted(num nodes/num nodes )
Click Finish and it ask you "Do you want to perform another operation ?" Choose Yes
Back to the operations screen where you select database .
on the node selection choose select all...
defined a db template- General purpose
on the management options enter the database name and the sid prefix.
then on step 6 of 16 use the same password for all account
On the storage option screen . select Automatic storage management -
select the DATA diskgroup to be used for the db storage . uncheck FRA
On the database files location screen . select oracle-managed file
Select the specify flash recovery area check box and enter FRA diskgroup to be used for fla
select some sample schema if you please
Create database service if you like ( to differentiate based on usage , batch for 1 node,oltp for both
then click , click
on the Creation option page select the create database option .
then click click click and you got a window
Starting cluster database "DEVL" and its instances "DEVL1,DEVL2 " in progress. ...
and now < I am done with my First RAC >
Subscribe to:
Posts (Atom)