I inherited a production system with the following setup:
- Disk A, Failgroup AB
- Disk B, Failgroup AB
- Disk C, Failgroup CD
- Disk D, Failgroup CD
Diskgroups AB and CD are configured with normal redundancy.
All disks are the same size. A and B share a common failure point, that is, it'd be unsafe for physical config reasons to make B a redundant version of A. Same for disks C and D.
If A and B's common failure point occurs and I lose disks A and B, what will happen to my database instance?
I've heard conflicting things:
- ASM will halt the database because it can't maintain it's redundancy configuration.
- ASM will attempt to use C and D and somehow rebalance the data from A and B.
- ASM will issue an alert and continue to allow things to run.
So what will happen? Is the result configurable?