ASM redundancy confusion

Question

Dear All,
o/s:redhat linux 5.6
db:10.2.0.3 with ASMlib
I always thought that ASM redundancy NORMAL will have only 2 failure groups and HIGH will have 3 failure groups.Is this sentence right or wrong because i say this below SQL statement on some website
SQL> create diskgroup DATA_NRML normal redundancy
Failure group flgrp1 disk
‘/dev/rdsk/c3t19d3s4’,‘/dev/rdsk/c3t19d4s4’,‘/dev/rdsk/c3t19d5s4’,
‘/dev/rdsk/c3t19d6s4’
Failure group flgrp2 disk
‘/dev/rdsk/c4t20d3s4’,‘/dev/rdsk/c4t20d4s4’,‘/dev/rdsk/c4t20d5s4’,
‘/dev/rdsk/c4t19ds4’
Failure group flgrp3 disk
/dev/rdsk/c5t21d3s4’,‘/dev/rdsk/c5t21d4s4’,‘/dev/rdsk/c5t21d5s4’,
‘/dev/rdsk/c5t21ds4’
Failure group flgrp4 disk
/dev/rdsk/c6t22d3s4’,‘/dev/rdsk/c6t22d4s4’,‘/dev/rdsk/c6t22d5s4’,
‘/dev/rdsk/c6t22ds4’;
if the above sql statement is correct then how normal redundancy works.
If failure group 1 had disk failure then mirror extents are on failure group 2 which will allow db to run.
If failure group 2 also have disk failure then db will be still up?
Regards

Answer 1

Hi,
-Can there be 3 failure groups in NORMAL redundancy? if yes then 2 diskgroup failure can be tolerated?No.
eg:
FAILGROUP controller1 DISK
   '/devices/diska1','/devices/diska2','/devices/diska3','/devices/diska4'
  FAILGROUP controller2 DISK
   '/devices/diskb1','/devices/diskb2','/devices/diskb3','/devices/diskb4'
FAILGROUP controller3 DISK
   '/devices/diskc1','/devices/diskc2','/devices/diskc3','/devices/diskc4'In this case with normal redundancy is tolerated a loss of one or all disk of one Failgroup at time. If fails disk from two or more failgroups at same time the diskgroup will fail.
see this example:
{message:id=10185895}
-If the 2 diskgroup failure cannot be tolerated then what is the use of adding 3rd failure group?we will go by 2.When I must choose of use different failgroup? When all disk(LUN) on that failgroup share same point failure. eg. (same array, same controller, same storage and so on).
Useful doc:
Use a Simple Disk and Disk Group Configuration
http://docs.oracle.com/cd/E11882_01/server.112/e10803/config_storage.htm#CDEEFBHG
Use Redundancy to Protect from Disk Failure
http://docs.oracle.com/cd/E11882_01/server.112/e10803/config_storage.htm#CDEBCDBD
Regards,
Levi Pereira