> Andrew Sullivan's fairly regular response is that he tried (albeit
> not VASTLY extensively) to distinguish between disks when working
> with fibre arrays, and he couldn't measure an improvement in
> shifting WAL (the OBVIOUS thing to shift) to separate disks.
Real quick... the faster the drives, the less important it is to move
WAL onto a different drive. The slower the drives, the more important
this is... which is why this isn't as necessary (if at all) for large
production environments.
-sc
--
Sean Chittenden