I'm now confused on how to physically setup my system. Based on these components can someone tell their view on what is the best correct way to set up an medium size oracle server with the proper fault tolerance levels and please be detailed. Meaning what RAID level(s), partitions and even about the paths for the control files and redo logs and how many of each.
COMPONENTS: RAID CONTROLER: 1 or 2 CHANNELS
HOT SWAPABLE SCSI DISKS: 3-4
RED HAT LINUX (can be any UNIX though)
Thanks in advance. Really appreciate it.
I typically setup each redo log group with three members on different drives. Put datafiles and index datafiles on separate drives, and try to separate high i/o datafiles so no two are on the same drive. All of this is in an effort to prevent contention and eliminate waits in the database due to high I/O. RAID5 is what many people like to use, but in the case of Oracle it severely limits what you can do to prevent/reduce contention since the data is striped across the drives. Mirroring drives is probably the best bet.
Click Here to Expand Forum to Full Width