29 | 03 | 2024
Latest Articles
Popular Articles

Administration

ASM 11G enhancements

User Rating:  / 0
PoorBest 

ASM 11G enhancements

 

Some of the enhancements are

 

1. new os role SYSASM : separation between SYSDBA and ASM administrator

2. ASM_PREFERRED_READ_FAILURE_GROUPS : with 10G ASM always reads the primary copy of an extent. with 11G - optionally - the ASM administrator can configure the ASM_PREFERRED_READ_FAILURE_GROUPS to assure that extent closest to the local node ( RAC ) are preferred.

3. compatible.asm and compatible.rdbms diskgroup attribute.

3.1 COMPATIBLE.ASM compatibility level is an attribute for an ASM diskgroup. 1 ASM instance can contain multiple diskgroups. Diskgoup +DG1 can have compatible.asm=11.1.0.0.0 whereas diskgroup +DG2 can have compatible.asm=10.1.0.0.0.

One of the new  ASM features is that the diskgroup allocation unit size - extent size - can be specified at diskgroup level. Also when a database file grows, the extent size will  automatically increase ( read automatically bigger allocation unit size is chosen ) . Large database files built in an Oracle 11G asm disk group have thus fewer extents than with Oracle 10G asm disk group, which lowers the amount of memory needed from the shared pool.  

This requires that compatible.asm=11.1.0.0.0 since this attribute could not be specified for a 10G ASM diskgroup. 

3.2 COMPATIBLE.RDBMS compatibility level of RDBMS instance. Example if COMPATIBLE.RDBMS = '11.1' is set at level of a diskgroup then no Oracle 10G database can use these diskgroups. 

ASM diskgroup compatible.rdbms must be lower than or equal to ASM diskgroup compatible.asm.

ASM diskgroup compatible.rdbms must be lower than or equal to DB instance compatible parameter.

ASM diskgroup compatible.asm must be lower than or equal to ASM instance compatible parameter.

 

4. There is a new view v$asm_attribute

5. FAST MIRROR RESYNC. After failure of an ASM disk with compatible.asm 11.1 or higher, the disk is taken offline but NOT immediatly dropped as it was with 10G. The asm disk is only dropped after a DISK_REPAIR_TIME amount of time has been exceeded. DISK_REPAIR_TIME can be set at level of a diskgroup to allow a period for stale modified extent tracking. If not repaired within DISK_REPAIR_TIME then fast mirror resync is not possible anymore, the disk is dropped and must be added again afterwards ( 10G functionality ). Fast mirror resync allows extents to become stale only when they are modified and only those are resynced once the disk is repaired, thus resync is faster

7. Bad Block Repair : runs automatically in background ( with normal and high redundancy disk groups) .

8. MOUNT a diskgroup in RESTRICTED mode

9. MOUNT a diskgroup with FORCE option

10. DROP diskgroup with FORCE option

11. new ASMCMD commands : cp, md_backup, md_restore, lsdsk, remap