Database Reference
In-Depth Information
Checking for Logical Corruption in ASM disk groups
The alter diskgroup command has been extended to include the scrub clause. This clause allows the ASM
administrator to check ASM diskgroups, individual disks in a disk group, or a single ASM files for logical corruption in
cases where ASM is responsible for protecting the data. If logical corruption is detected during the scrubbing process
ASM can try to repair it using mirror copies of the extent. By default ASM will only report corruption, you need to
specify the repair keyword to let it try to fix it.
The final relevant parameter for this operation is the power level. Unlike some other ASM commands you do not
specify an integer, but rather any one of low, high, and max.
If you suspect a problem with the diskgroup “RECO” of your system, and want to be reported of potential
problems rather than fixing them on the spot, you could use the following command:
SQL> alter diskgroup reco scrub power high;
Diskgroup altered.
No further information was found to be returned to the prompt, but it is expected that a summary of errors will
be returned if there are any. Just like the check diskgroup clause additional information is recorded in the ASM
instances alert.log. In the above example, which found no problem the alert.log information contained these
few lines:
2013-10-10 08:59:42.373000 +01:00
SQL> alter diskgroup reco scrub power high
NOTE: Start scrubbing diskgroup RECO
Starting background process SCRB
SCRB started with pid=34, OS id=7934
SUCCESS: alter diskgroup reco scrub power high
Scrubbing will be prevented if the I/O load of the ASM instance is currently considered too high. You can view the
progress of the scrubbing operation in v$asm_operation as usual.
ASM Becomes More Flexible
Before Oracle 12c every node in the cluster needed to have its own ASM instance. This collection of ASM instances is
generically referred to as an ASM cluster. Instances within the cluster communicated with each other with the goal to
present shared storage to all users. Up until Oracle 11.2 this included databases only, but with the introduction of the
general purpose file system based on ASM, named ASM Cluster File System or ACFS for short Oracle offers a complete
stack of a cluster aware storage solution.
Up until 12c it has been the case that a failed ASM instance meant all the databases on the host were to fail as
well. Oracle Flex ASM, as the new set of features has been named, addresses this situation by removing the strict
requirement to have one ASM instance per cluster node. The default configuration of ASM up to version 12c is shown
in Figure 2-4 .
 
Search WWH ::




Custom Search