ASM disk group administration-Creating a disk group

In general, once an ASM instance completes the ASM disk's discovery operation, you can then use the ASM disks to either build a disk group or add it to a pre-existing disk group. You need at least one eligible ASM disk (with HEADER_STATUS either
CANDIDATE, FORMER, or PROVISIONED) to be able to build a disk group or to add to a pre-existing disk group. There are various methods to create a disk group, such as, DBCA (in pre 11g R2), ASMCA (with 11g R2), Grid Control, and CREATE DISKGROUP SQL statement in the SQLPLUS prompt. Although the easiest and most convenient way to build and manage a disk group is to use the GUI tools such as DBCA, ASMCA, or Grid Control, we are going to demonstrate how to create and manage a disk group using a set of SQL statements.

In order to create a disk group on the local ASM instance, you first need to identify the eligible ASM disks discovered by the instance. The following SQL statements are useful to list the eligible ASM disks:

export ORACLE_SID=+ASM1
export ORACLE_HOME=$ASM_HOME (in 11gR1)
export ORACLE_HOME=$ORA_GRID_HOME (in 11gR2)

SQL> SELECT path,header_status,mode_status,total_mb
FROM v$asm_disk WHERE header_status IN ('CANDIDATE','FORMER','PROVISIONED');

The query lists each disk's information along with the path, the header status, and size of the disk. Once you list the disks, you can use the following set of SQL statements to create a new disk group with different levels of mirroring options.

The following SQL statement creates a new disk group named DATA with one ASM disk, sde1 located under the /dev/ location. The EXTERNAL REDUNDANCY clause in the SQL statement indicates that you are relying on the STORAGE-level mirroring
(protection) option, not using the Oracle-provided mirroring level:

CREATE DISKGROUP data EXTERNAL REDUNDANCY DISK '/dev/sde1';

The following SQL statement creates a new disk group named DATA that consists of two failure groups with one disk to each failure group, using the Oracle-provided NORMAL redundancy (a two-way mirroring) level:

CREATE DISKGROUP data NORMAL REDUNDANCY FAILGROUP fgp1 DISK '/dev/sde1', FAILGROUP fgp2 DISK '/dev/sdf1';

The following SQL statement creates a new disk group named DATA with three failure groups with one disk to each failure group, using the Oracle-provided highest level of redundancy, a three-way mirroring-level option:

CREATE DISKGROUP data HIGH REDUNDANCY
FAILGROUP fgp1 DISK '/dev/sde1', FAILGROUP fgp2 DISK '/dev/sdf1', FAILGROUP fgp3 DISK '/dev/sdg1';

Note: Failure groups are used to copy the redundant copies of every extent. When the mirroring option is skipped, ASM applies the NORMAL REDUNDANCY (a two-way mirroring) level by default.

After a disk group is successfully built, it will then automatically mount in the local instance. If a server parameter file is being used (SPFILE), the name of the disk group is successfully added to the ASM_DISKGROUP initialization parameter in order
to mount the disk group automatically on ASM instance restarts. In order to make the disk group available on other ASM instances in a cluster, you simply need to mount the disk group running the ALTER DISKGROUP data MOUNT statement (ensure the same set of disks are accessible on the other nodes).