Oracle 11.2 allows the DBA to define the block size for redo logs:


http://tinyurl.com/6co7u2e

The DBA guide is mostly concerned with the redo wastage. The available
options are 512B, 1K and 4K. The 4K size is recommended for the newer
disks, with 4K sector. On Linux, however, the vast majority of file
systems use 4K blocks which leads me to the conclusion that there would
be better blocking factor for large ETL type operations with 4K redo log
block size. Did anyone experiment with that? Are there any experiences
whatsoever?

--
http://mgogala.byethost5.com

On Nov 2, 1:12*pm, Mladen Gogala <gogala.mla... (AT) gmail (DOT) com> wrote:
Of course, I have no experience and don't know much about it, but the
first thing that strikes me is, wouldn't it be based on the disk
itself, whether it uses emulation mode? I'm thinking this option
would be a way of getting around filesystems hiding it from Oracle.

jg
--
@home.com is bogus.
People want to port vms to x86?
http://searchoracle.techtarget.com/n...rten-user-says