Datasheet
18
Chapter 1
Using Oracle ASM
You can change the amount of time that Oracle will wait to automatically drop the disk by
setting the
disk_repair_time attribute (see the discussion on attributes earlier, in the section
“ASM Disk group Attributes”) for the individual disk groups using the
alter diskgroup com-
mand, as shown in this example, where we set the
disk_repair_time attribute to 18 hours:
Alter diskgroup dgroup1 set attribute ‘disk_repair_time’=’18h’;
ASM Preferred Mirror Read
The ASM preferred mirror read feature allows you to define a primary set of disks that are
the preferred disks to read from for a given instance. This is most prevalent when using
RAC databases. In a RAC configuration, you could have two or more sets of disk arrays.
Each disk array might be local to a given RAC instance. ASM preferred mirror read allows
you to indicate which disk array is local to a specific RAC instance. As a result, it become
the preferred disk set for the instance and thus is likely to be more performant.
The ASM preferred local disk is defined using the optional parameter
asm_preferred_
read_failure_groups
.
ASM AU Size and Extents
ASM files are stored in disk groups. In each disk group, space is allocated in extents, and an
extent consists of one or more units of space called allocation units (AUs). Allocation units
default to a size of 10MB and can be configured from 1 to 64MB at the time the disk group is
created. Once the AU size has been determined for a given disk group, it cannot be changed.
To enable support for larger ASM datafiles, to reduce the memory overhead of large
databases, and to improve file open and close operations, ASM uses a variable-extent sizing
policy. Extents will be initially sized at the size of the AU (for the first 20,000 extents) of the
ASM disk group in which the extent is created. The extent size will be increased to 8 times
the AU size for the next 20,000 extents and then will increment to 64 times the AU size for
subsequent extent allocations.
Adding an ASM Disk Group
We have now talked about discovering disks, and we have talked about templates, redundancy,
and striping. Now we need to talk about actually creating a disk group. You use the
create
diskgroup
command to create an ASM disk group. When you issue the command, you will
assign the disk group its name, and you will add one or more discovered (unallocated) disks to
that disk group. Here is an example of the use of the
create diskgroup command:
CREATE DISKGROUP dgroup1 NORMAL REDUNDANCY
failgroup diskcontrol1 DISK
‘/devices/diska1’
failgroup diskcontrol2 DISK
‘/devices/diskb1’;
95134c01.indd 18 1/28/09 9:43:42 AM