This content has been marked as final. Show 2 replies
I think for our systems we typically do one instance with multiple disk groups per database. ie.
This makes things a lot easier to manage, especially if you get rid of a DB and don't want it to affect the others.
Take a look at http://docs.oracle.com/cd/E11882_01/server.112/e18951/asminst.htm#BHCFHDDJ
this gives you a formula to calculate PROCESSES parameter.
There can be only one ASM instance per node, so if you have many database instances running on one node they must share the same ASM instance and there is no problem with that.
Note "ASM Technical Best Practices [ID 265633.1]" states that "Oracle generally recommends no more than two diskgroups." So different databases can share common disk groups easily and it makes administration tasks much easier. I have usually created additional TEMP diskgroup for storing tempfiles, because tempfiles could be excluded from storage replication to standby site.