I am planning on setting up a 11gR2 RAC environment using SAN with hardware raid redundancy for the ASM disks. I've been looking at different blogs, white papers and docs and am little unclear on a couple points.
1) Should I store the voting disks/ocr in a separate disk group than the +DATA disk group? Behind the scenes both disk groups would end up using different LUNs from the same SAN server.
2) Using external redundancy for the voting disks results in the creation of a single voting disk. Is there any harm to having a single voting disk when using external hardware redundancy? Or should I create 3 disks and assign them to the voting disks/ocr diskgroup and set the redundancy to normal so that 3 voting disks get created? I am just unclear if this would result in any benefit given that I have external hardware redundancy.
i think you should use a dedidcated diskgroup for the cluster like DG_CLUSTER with other den external redundancy. You also mirror your control files or your redo log files, why not mirror your ocr or your voting files?
You can also change your redundancy level (more than one failgroup) with one storage in the backend.
You can use another Diskgroup than DATA, like "OCR" or something like that. It doesnt matter if you are using external (sotrage) or normal (ASM) redundancy, as long as you use redundancy... that is important... You can use a diskgroup with a little space to store your OCR files.
Control Files redundancy is good Idea as mentioned on last post....