The log buffer is flushed when it reaches 1/3 full, or it reaches 1M
(ignoring the 3 second rule for the moment), so it can never possibly
hold 8M. Anything over 3M is inherently wasted space. That said, we
have always found smaller sizes than that to be fine here. Depends on
your environment of course, you have to set it and check to make sure
that you are not seeing waits for log buffer space.

HTH

-----Original Message-----
From: James Oddo [mailto:[Email Address Removed]
Sent: Tuesday, April 20, 2004 1:59 PM
To: LazyDBA.com Discussion
Subject: Oracle 9i memory allocation on 2000


Does 8MB for redo_log_buffer seem excessive
Also
db_block_size = 16384
db_file_multiblock_read_count = 64
db_keep_cache_size = 150M
db_cache_size = 900M
java_pool_size = 64M
shared_pool_reserved_size = 32M
shared_pool_size = 160M


At what point does too much memory become a hinderance if any


Thanks
Oracle LazyDBA home page