Re: v$filestat

This is only for direct writes from sorts, index builds, direct loads, CTAS, etc. which reminds me that hash_multiblock_io_count is good for multi-block writes too.

Any writes that go through dbwr/cache
are single block writes.

--
Jonathan Lewis
Yet another Oracle-related web site:  http://www.jlcomp.demon.co.uk

Practical Oracle 8i:  Building Efficient Databases
Publishers:  Addison-Wesley

Reviews at: http://www.jlcomp.demon.co.uk/book_rev.html



Nuno Souto wrote in message <3aa75761.4759524_at_news-server>...


>On Thu, 8 Mar 2001 09:23:21 -0000, "Jonathan Lewis"

><jonathan_at_jlcomp.demon.co.uk> wrote:

>

>Thanks, Jonathan.

>Is that one reliable and consistent with the ratio of writes to block

>writes?  Ie, does dbwr actually sequence blocks from the cache to make

>them write sequentially in a single command?  Or is it just a multiple

>write command initiation, no actual physical sequence involved?

>

>>

>>db_file_direct_io_count ?

>>

>>(only in 8+)

>>

>>--

>>Jonathan Lewis

>>Yet another Oracle-related web site:  http://www.jlcomp.demon.co.uk