I assume the partitioning is filesystem level. So for me I used to split out the redo logs to raid 0 only and data to raid 0+1 (10) and indexes to the same. If you are running with archive logs then split them off as well. Also binaries on their own separate mount point. Or to make life a little easier in the long run use ASM. I hated it to no end in the beginning but since we have been using it I have grown to appreciate it's simplicity. We use ASM and OMF with bigfiles so IO performance has never been an issue except with bad sql and file management is no longer an issue at all. Still you will have to split off some diskgroups for archive logs. Also ASM doesn't look after the binary installation mountpoint. But all DB files control files, redo logs, and datafiles are very well taken care of.
Last Wiki Answer Submitted: January 19, 2006 11:46 am by Sequential0 pts.