Home > Workload Solutions > Oracle > White Papers > Oracle ASM on ScaleIO Best Practices > ASM redundancy
ASM includes the ability to mirror diskgroups to protect data from physical failures.
Available protection levels are NORMAL (one mirror), HIGH (two mirrors) or EXTERNAL where the data protection is handled by ScaleIO.
ScaleIO automatically protects the ASM data using a MESH mirror technique, whereby every block of data is held on two different nodes of the ScaleIO system. In the event of failure, ScaleIO immediately recreates a redundant copy of the data on a surviving node.
Therefore protecting data with ASM is unnecessary and ASM diskgroup protection be set to EXTERNAL.
Note: When storing cluster voting disks in ASM, the DBA may consider using a protection level of NORMAL with two failure groups.
This will create mirrored copies of the voting disks as shown in Table 5:
Table 5. Voting disks and failure groups
ASM redundancy |
Copies of voting disk |
Number of failure groups required |
External |
1 |
N/A |
Normal |
3 |
2 |
High |
5 |
3 |
See Oracle documentation Managing Oracle Cluster Registry and Oracle Local Registry sub-section Storing Voting Disks on Oracle ASM.
Dell EMC recommends using external redundancy in all ASM diskgroups except for ASM diskgroups where cluster voting disks are stored, for which normal redundancy with two failure groups should be considered.