Database Software

Hello,

I would like to verify that I correctly understand the concept of ASM Failure Groups.

According to Oracle documentation, a Failure Group is a set of disks whose failure needs to be tolerated. ASM assigns each disk to it's own failure group, by default.

Question A:

From what I understand, ASM will mirror the content of disks or a group of disks only provided it is a "Failure Group". Data redundancy under ASM is either 2-way or 3 way and works on the level of file extents. In other words, if a disk or group of disks is not a Failure Group, then ASM will not provide data redundancy. Is this correct?

Queston B:

If the above is correct, what is the purpose of using a set of disks as a Failure Group? Does this not mean that the failure of a any member or any file extent in the Failure Group will be considered a failure of the whole disk group, and make the disk group offline, even if just a single disk failed? Why would I want this? Is this necessary to provide data redundancy between different SCSI channels, as otherwise the loss of a SCSI channel might cause the loss of disks needed for redundancy?

Thanks.

Edited by: Dude on Nov 27, 2010 2:53 PM