Simple and precise
Take analogy of RAD 1+0 with 4 disks. 2 disks (on controller 1) are in one failure group and 2 disks(on controller 2) are in second failure group. If controller 1 fails and both disks are gone, your other failure group is still available and so is your data because controller 2 disks are a mirror of control 1 disks.
Failure groups are used to determine in which ASM disks can be used for storing redundant copies of data. Failure groups are applicable only in 2way mirrorning(normal redundancy), 3 way mirroring(high redundancy). so that if one disk group failed, the other group can help the node be survived still.
asm_power_limit parameter specifies the maximum power for disk rebalancing. it can be set from 1 to 9. The more you set the disk rebalancing will be much effective and also consumes more processing and IO resource. The less value will consume less processing and IO resource. the default value for this parameter is 1
Hope this answered..
Failure group is a logical subset of an ASM diskgroup in case of Normal/High Redundancy diskgroup.
Each failgroup content subset of disks which physical entity of failgroup.
Failgroup is used to keep mirror data ,that means ,each disk on one failgroup make partnership to another disks on other failgroup,cannot be from same failgroup.
So,a allocation unit placed on that disk will have mirror on another disk which partner with that.
asm_power_limit controls how many slave process (ARBnnnn) will spawn while you do diskgroup reconfiguration/explicit diskgroup rebalance .
Default value is 1 .if you set to 0 then rebalance will not take place automatically.
You can set this value in the range of 1 - 1024.