A database instance is using an Automatic Storage Management (ASM) instance, which has
a disk group, DGROUP1, created as follows:
SQL> CREATE DISKGROUP dgroup1 NORMAL REDUNDANCY
FAILGROUP controller1 DISK ‘/devices/diska1’, ‘/devices/diska2’
FAILGROUP controller2 DISK ‘/devices/diskb1’, ‘/devices/diskb2’;
What happens when the whole CONTROLLER1 Failure group is damaged?
A.
The transactions that use the disk group will halt.
B.
The mirroring of allocation units occurs within the CONTROLLER2 failure group.
C.
The data in the CONTROLLER1 failure group is shifted to the CONTROLLER2 failure
group and implicit rebalancing is triggered.
D.
The ASM does not mirror any data and newly allocated primary allocation units (AU) are
stored in the CONTROLLER2 failure group.
I do not think asm will rebalance the functional diskgroup in this instance. In this scenario the redundancy level is set to normal. This means there will be two-way-mirroring; one on failure group 1 and one on 2. If the entire failure group 1 fails no data will be moved to 2 so no rebalancing. Does anyone agree with this theory?
After the failure, the DG is left without mirrors. So after the drop of the diskgroup is complete (after a failgroup failure its disks are put offline), implicit rebalancing is started to guarantee the mirroring level
Answer C
see
https://docs.oracle.com/cd/E11882_01/server.112/e18951/asmdiskgrps.htm#BABJIBII
although how it is put in answer C here is somewhat ambiguous.
I changed my mind, the correct answer is D.
@Andreas, above is correct. Mirrowing occurs between failour groups and not within failour groups. I searched the documentation and found nothing that even suuggests otherwise.
however I did find that when you create a NORMAL disk group it is recommended that you give it 3 failour groups. Note in the question above we use 2 (which is the min). If there were 3 then Oracle could then mirror to the 3rd failour group, so you would have all files replicated to failour groups 1 and 3 which would still give you NORMAL redundancy.
@Noa, the disk group remains online after the failour, this gives us availability even though one of the failour groups has failed. Hey, after all, this is why we pay so much for this technology and for the hardware right!!! We are in effect buying “availability” and the disk group is not droped.
I stick on C. because rebalance may happen if a disk of failure group is failed.
https://docs.oracle.com/cd/E11882_01/server.112/e18951/asmdiskgrps.htm#BABJIBII
“After the disk drop operation, Oracle ASM performs a rebalance to restore full redundancy for the data on the failed disks.”