Hey Ted, I knew I hadn't stated that very well. I never thought that any single extent would be striped among datafiles, since the definition of extent is "contiguous blocks". Since I'm newish to RMAN and am far from a storage expert, I was fishing for something like "never expand old datafiles because...". Also, since I arrived here one month prior to the final implementation phase of this ERP, ASM was out of the question. :) It seems like I was happily barking up a nonexistant tree. Hopefully there was no one in the forest to hear it, or some other nonsensical metaphor... Thanks, Ted! Rich > Extents can't span datafiles. Each extent will be used across datafiles in > round robin fashion. > > Example: 1 tablespace, 2 files and one table with two extents. The first > write is to extent one file one, the second write is to extent two file two, > then back to one and so on. Possibility exists that the first file has more > extents based on initial extent allocation, but I didn't test this out. > > This is light on the technical, but it might help. > > Personally, I'd use fewer, larger files, but that point is based on the > arbitrary nature of the installation, meaning fewer larger files work in > some cases, and not in others. I'd prefer to go ASM bigfile on everything > and not worry about storage allocation, but no luck with that yet. > > -T -- //www.freelists.org/webpage/oracle-l