KEEF-01	Ensure that all 3 control files are on different disks (as long as we have >= 3 disks).
KEEF-02	Keep DATA and INDEX on different disks.
KEEF-03	Break NH_DATA01 across several disks (even when NOT hitting the 32G file limit)
KEEF-04	Break NH_DATA02 across several disks (even when NOT hitting the 32G file limit). Not sure if we care that NH_DATA01 and NH_DATA02 are on the same disk.
KEEF-05	Break NH_INDEX across several disks (as long as we have enough disks to not violate KEEF-02)
KEEF-06	Keep TEMP away from DATA and INDEX (if possible)
KEEF-07	Ignore where NH_HOME and ORACLE_HOME are
KEEF-08	Do something much smarter with REDO LOGS.
KEATS-001	We should have 4 disks for redo logs as follows:
 
    disk 1: redo1a, redo3a
    disk 2: redo1b, redo3b: 
 
    disk 3: redo2a, redo4a
    disk 4: redo2b, redo4b
 
KEATS-002	Archive logs should be on different disk than any redo logs
KEATS-003	What happened to NH_USERS?  OK it's not a "rule" but it's not in the a-i block.
KEATS-004	Redo logs and Archive logs should not share a disk with tablespaces (they'd be better off buying some cheap tiny drives (2GB) for the Redo logs)
KEATS-005	Keep NH_USERS separate from the DATA and INDEX disks
KEATS-006	Keep INDEX on it's own disk
KEATS-007	Definitely keep NH_TEMP on it's own.
KEATS-008	I was told by Sanjay to keep ORACLE_HOME on a different disk from the DB.  I've done this in performance testing whenever possible.
KEATS-009	What out for rule "KEEF-01"...what happens if you're striping to one logical volume?  Does createDb handle this?
KEATS-010	How do we handle striped disks?
KEATS-011	I'll tell you later. :)
KEATS-012	you should treat FE and BE differently...on a FE the DATA 01/2 tablespaces are unused, so if you keep the same rules for both you will put these on some of the largest disks.
KEATS-013	how do you handle the situation where a customer has a bunch of 18GB disks but needs a 20GB DATA01 datafile?
KEATS-014	if you are going to start splitting datafiles (even before the 32 GB limit) b/c of "extra" disks, then I'd recommend splitting them one at a time (just INDEX and DATA01/2) instead of breaking one into many pieces first.  For example, if you have 4 "extra" disks, then split INDEX and DATA01/2 into 2 pieces first...then make the biggest of the three into 3 pieces.  Of course, I would do this AFTER everything else is on its own disk.
KEVILLE-01	Check during db creation that the stripe width matches our db block size, if they are using striping, if not ask if they want to continue.
KEVILLE-02	Check the spindle speeds of the disks they are installing on and try to match tablespaces that cross disks to same speed disks(?)
RAVI-01	System, redo and archive locations should be on different non stripped individual disk if possible. They should not be place along with other files if possible. Also they should be placed on the smallest files if possible 
RAVI-02	The two redo’s and the archive locations should be on different disks 
RAVI-03	The 3 control files should be on different disks 
RAVI-04	The index and data1 and data2 should be on different disks. (possibly even different raid arrays.) 
RAVI-05	We need to split tablespace into one or more files based on a max file limit that is specified depending on the tablespace using the below table. That is for the index tablespace this limit is 10GB. For temp it is 100MB etc. The reason for this rule is to allow us to exploit many individual disks and spread IO without the need for RAID. 
RAVI-06	We need to allocate the index, data1, data2 starting from the largest disk to the smallest disk in the order specified. 
RAVI-07	The system location must be on one and only one disk.