Database Reference
In-Depth Information
Replacing ASM Disk in Disk Group
Pre-12c, to replace an existing ASM disk in an ASM disk group for whatever reason, you had to add a new disk first and
then drop the existing disk. With the new REPLACE CLAUSE in 12c, you can easily replace an ASM disk in an ASM
disk group with just a single operation, in contrast to the previous add-and-remove disk procedure. The following
example demonstrates how to replace an ASM disk (ensure you connected to an ASM instance as sysasm):
SQL> ALTER DISKGROUP dg_data REPLACE DISK dg_data_004 WITH '<new_asm_disk>' POWER 4;
Scrubbing ASM Disk Groups and Files
The new ASM SCRUB feature in 12c got the potential to verify any logical data corruption on any ASM disks, disk
groups, or files and optionally repair them using ASM mirror disks when an ASM disk group is configured in either
a high or normal redundancy level. The disk-scrubbing operation can be performed with the combination of HIGH,
MAX, LOW, and/or AUTO option to control the I/O on the system, and this sort of operation typically will have very
little I/O impact on the server.
The following examples demonstrate how to execute the command to verify logical corruptions as well as repair
them automatically:
SQL> ALTER DISKGROUP dg_data SCRUB POWER AUTO:LOW:MAX:HIGH;
-- Default POWER option is set to AUTO
The following example runs through logical data corruption on the data file and repairs any corruption using
the mirror copy:
SQL> ALTER DISKGROUP dg_data SCRUB FILE '+DG_DATA/MYDB/DATAFILE/users.374.817049597'
REPAIR POWER HIGH;
If no REPAIR clause is specified, then only logical data corruption verification is performed. You can monitor
the progress of the procedure in the V$ASM_OPERATION dynamic view.
Reading Data Evenly in ASM Disk Group
A new feature in 11g, Preferred Read Failure Groups, provides the functionality to read the data from an extent which
is close to the node to improve the read efficiency for a disk group configured with a normal or high redundancy level.
A new default feature in 12c, Even Read for Disk Groups, provides the ability to distribute the data read operation call
evenly across all disks in a disk group. When a disk group is configured with a normal or high redundancy level,
each read request can possibly be sent to the least-loaded source ASM disk to improve the read efficiency.
Measure and Tune Rebalance Operation
To measure any ASM disk group rebalance operation in order to tune or control the overall duration required for the
rebalance operation, you can use the new EXPLAIN WORK FOR statement on AMS instance. The new SQL statement
calculates the estimated work required for the rebalancing operation and puts the result in the V$ASM_ESTIMATE
 
Search WWH ::




Custom Search