Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: Common mistakes when upgrading to 9i

Re: Common mistakes when upgrading to 9i

From: Howard J. Rogers <hjr_at_dizwell.com>
Date: 26 Feb 2004 09:45:00 -0800
Message-ID: <14a1f766.0402260945.6a97ce01@posting.google.com>


netcomrade_at_karlson.dmitriev.net (NetComrade) wrote in message news:<slrnc3pr5b.t17.netcomrade_at_karlson.dmitriev.net>...
> Hi,
>
> We are finally on a path to upgrade from 8i (8.1.7.4) to 9i (9.2.0.1).
> I wanted to learn from your experiences on things that can go wrong.
> I am going to list a few things I've read up on, and if you're kind enough,
> you'll let me know if I missed anything important
>
> * sga_max_size should be examied
> * pga_aggregate_target default is 25MB, if set DB ignores
> sort_area_size
> hash_area_size
> bitmap_merge_area_size
>
> How do above two affect MTS (Shared Server)
>
> * workarea_size_policy defaults to auto if pga_aggregate_target is set
> * db_cache_size replaces db_block_buffers and is in bytes
> * statistics_level defaults to typical
> * automatic undo management should be researched (but not req'd)
> * db_block_checksum is true by default in 9.2 (?)

The commonest mistake I've seen made is one that has been discussed here just recently: the default cache and the recycle cache and the keep cache are now three separate entries whose sizes need to be added up to give you your total buffer cache size, whereas before the keep and recycle caches were merely subtracted away from a total cache size set by db_block_buffers. If you literally translate block_buffers into db_cache_size, buffer_pool_keep into db_keep_cache_size and buffer_pool_recycle into db_recycle_cache_size, then you're in for a shock because your memory consumption is going to go through the roof.

Regards
HJR Received on Thu Feb 26 2004 - 11:45:00 CST

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US