Re: Viewing the hidden _{Parameters}

My first thought was that this was perhaps one of those odd 'boundary limit' calculations, or a platform dependent feature, so I've bounced by HP with a 36,000 block buffer and started to dump the buffer chains.

I've just realised though that you meant the x$ksppi query, not the 'alter session' "query".

There are several places where Oracle takes a supplied init.ora parameter, and then adjusts it to suit. In this case I think we'll find that the parameter _db_hash_buckets is first generated as a simple multiple of db_block_buffers, and then adjusted upward to a prime when the instance starts. I'll let you know when the dump completes - it's having a bit of a problem because of the number of buffers, and the fact that there seems to be a bug that increases the dump size whenever Oracle hits an unused buffer (and since I've just bounced the database ...)

--
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



Rob K wrote in message ...


>I did a check using your query and my _db_block_hash_buckets  = 72000

>which = 2*db_block_buffers = 36000

>not the next prime of 72000

>which is 72019

>

>I am running 8.1.7.0 on Sun

>I heard there is a patch release to bring me

>up to 8.1.7.1

>

>