I am about to migrate multiple Oracle instances from a Solaris OS/EMC RAID 5 to an AIX p575 (two of them) with 50TB of Data Direct storage. As a self-taught Oracle admin of three years, I may not be doing things the best way now, and before I migrate I would like to have a good plan for laying out my instances. Currently I have my various instances setup as follows:
In this configuration, all of the d0* are separate filesystems. I have liked this setup, because all of my tablespaces reside under the appropriate SID. It is easy to take care of from a Unix perspective and Oracle perspective, except for one thing. If I misjudge the size needed for a given SID's d0*s, then I am stuck with unused space.
Can you recommend an alternative? Thanks.
One alternative is to upgrade to Oracle 10g and use Automatic Storage Management (ASM). With ASM, you let Oracle be the volume manager for you and have it take care of these details.
If you do not want to use ASM, then consider the following scheme for your mount points:
In the above, I have two mount points, /opt/oracle/oradataXX, and in each mount point, I create a subdirectory for each SID. This way, multiple databases can share the same mount point so you do not waste space. The big thing you'll want to watch out for is that two databases do not share the same hot disk volume. Spread out your hot spots so that one database will not impact another.
Dig Deeper on Oracle database design and architecture
Have a question for an expert?
Please add a title for your question
Get answers from a TechTarget expert on whatever's puzzling you.