Re: Default init.ora (Was Re: Oracle vs O'Reilly Performance Tuning Books)

From: Paul M. Aoki <aoki_at_faerie.CS.Berkeley.EDU>
Date: 8 Oct 1999 01:36:51 GMT
Message-ID: <7tjhrj$h9v$1@agate-ether.berkeley.edu>

Kelly Burkhart <kburk_at_sky.net> writes:

| aoki_at_faerie.CS.Berkeley.EDU (Paul M. Aoki) writes:

| > the niemiec book also tells you a bunch of stuff that you DON'T
| > get from the oracle documentation and directly contradicts some
| > of the "tips" in the sample files (like the example init.ora,
| > which is just plain evil incarnate) and documentation. i found

| 
| What is "evil incarnate" about the init.ora?  (Those are your words I
| assume).  Is it just missing stuff that a properly planned database
| should have?  Or does it actually contain bogus configuration?

the latter. but it's not bogus in the sense that things won't *work* - they just won't work very well. for example:

given today's memory configurations,

	db_block_buffers = 100
	shared_pool_size = 3500000

seems somewhat unmotivated. at least they caveat this part. (don't laugh - when i took over our local installation the guy in charge had left those parameters in "because that's the default"..)

defaulting the block size to 2KB does not seem like a good idea, since (as they suggest in the documentation) the block size should be an integer multiple of the file system block size and most unix variants default to 8KB file system blocks. (maybe they try to use one init.ora for NT, unix, VMS... certainly this is suggested by the VMS file names in the archive log parameters!)

putting both control files in the same place seems highly unadvisable:

        control_files = (ora_control1, ora_control2)

to be fair:

i will say that using DBCA ('dbassist') was a much, much more positive experience (aside from the itty bitty redo log files). it also generated the standard $(ORACLE_HOME)/rdbms/admin loads that all you real oracle DBAs must know to do but somehow didn't make it into the "install" instructions in the Admin Guide...
--

  Paul M. Aoki         | Computer Science Division, EECS #1776
  aoki_at_CS.Berkeley.EDU | University of California
                       | Berkeley, CA 94720-1776 USA

Received on Thu Oct 07 1999 - 20:36:51 CDT