RE: does sga_target cause resizing of db_32k_cache_size

nods. I would have to dig out decades old VLDB notes to be sure, but if memory serves the group request was each possible pool should have an option for manual versus auto, but that choosing any manually sized pool should never make the rest of the automation impossible.  

mwf  

From: oracle-l-bounce_at_freelists.org [mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Jonathan Lewis Sent: Friday, March 12, 2021 6:21 AM
To: Oracle L
Subject: Re: does sga_target cause resizing of db_32k_cache_size    

It's occurred to me that in at least half the cases where I've used a non-default cache (whether variant size or keep/recycle) I would have been very annoyed if it had resized itself based on use. I've only done it rarely, but if I've had a non-default it's typically been to protect the default cache from one special object, so I wouldn't want the non-default to grow unless I requested it explicitly.  

Allowing a min and max for each cache (and other components) would be nice for those rare occasions, though.  

Regards

Jonathan Lewis    

On Thu, 11 Mar 2021 at 19:45, Mark W. Farnham <mwf_at_rsiz.com> wrote:

The sized ones (not default cache) was INTENTIONALLY left out of automation with extreme prejudice.  

--
http://www.freelists.org/webpage/oracle-l