Re: Optimal values in init.ora file

ejsanchez_at_gmail.com wrote in message news:<1101858641.436859.236580_at_f14g2000cwb.googlegroups.com>...
> Hi!
> I'm trying to optimize my init.ora file.
> We recently re-install this Oracle server(Redhat 9 + Oracle 8.1.7) and
> now it is too slow...before the re-install it was a fast work stallion!
> I lost my previous init.ora file :(
> I know there's plenty of information on how to set the SGA on Oracle
> but all the articles I already read aren't very clear for me.
>
> Here are my most important parameters in my init.ora file:
> open_cursors = 1000
> max_enabled_roles = 120
>
> db_files = 3000
> db_file_multiblock_read_count = 32
> db_block_buffers = 6400
>
> #Variable size
> shared_pool_size = 18000000
> large_pool_size = 32000000
> java_pool_size = 40000000
>
> log_checkpoint_interval = 10000
> log_checkpoint_timeout = 1800
>
> processes = 250
> log_buffer = 819200
> db_block_size = 8192
>
> This machine is a Pentium III with 256Mb of RAM, before the re-install
> it was
> pretty fast, but now we're having performance problems.
> Is there a simple formula to set this values? depending on how much RAM
> & Processor I have?
> thanks in advance for any suggestions
> regards,
> -eduardo s.m.

Rather than just take wild guesses about what your settings should be I think you should run a statspack report and look to see if any areas appear to be a problem.

The selected parameters you show indicate that you have a large pool allocated larger than the shared pool but you do not display any MTS or PQO parameters that would require use of the large pool parameter at other than the calculated value. Incomplete information often results in incorrect advice.

What kind of application do you have that needs a session to be able to hold a 1,000 open cursors?

Since you apparently rebuilt the database I hope you recalculated the object statistics.

HTH -- Mark D Powell -- Received on Tue Nov 30 2004 - 20:23:18 CST