ASM Fast Mirror Resync
In Oracle 10g, FAILURE,ASM will not be able to continue writing data to the ASM disk if it occurs because of a fiber, controller failure.
ASM will change the disk that occurs failure to offline, and then drop the disk after synchronizing the data from other redundant groups. ASM will no longer read data from the disk being offline, because ASM considers the disk's data to be outdated.
Even if the failure disk is automatically repaired, it is necessary to manually reseat the disk back into the resynchronization.
1. Offline the disk and specify the drop after statement
sql> alter diskgroup DGROUP1 disk DATA01 drop after 0h; #asm完成数据重新平衡后将该磁盘删除
2. Before re-joining the disk to the disk group, erase the data on the disk (this step is not required in 11g)
# dd If=/dev/zero of=asm_disk1 bs=1024k count=100
3. Rejoin the disk to the disk group
Sql> alter DiskGroup dgroup1 add disk '/dev/raw/raw1 ' size 100m;
In Oracle 11g, ASM Fast Mirror Resync reduces the cost of disk failure when it occurs in 10g.
If it is a transient disk, FAILURE,ASM will track modifications to the corresponding extent on the disk where the failure occurred, and when the disk resumes, resynchronize the modified extents.
This simply requires rewriting a small portion of the extents content. After the failure disk is offline, it will not be drop (need to see the settings for parameter disk_repair_time). The value of the parameter disk_repair_time indicates how long after the disk offline does not drop the disk.
Open asm Fast Mirror Resync
Turning on this feature requires setting the Disk_repair_time value for the disk group. Default is 3.6h
Sql> alter diskgroup Dgroup1 set attribute ' disk_repair_time ' = ' 3.6h ';
To join a disk group after failure disk maintenance is complete, simply do the following:
sql> alter DiskGroup DGROUP1 online; Data synchronization #自动使用ASM Fast Mirror Resync
Monitor ASM Fast Mirror Resync Process:
· V$asm_disk
· V$asm_disk_iostat
11G new Features-ASM Fast Mirror Resync