Re: Oracle on Windows Vs. Linux

Some obvious things to check, which link to the other questions.

Did you enable direct IO? Windows does this natively, you may - dependent on version of Oracle and your filesystem - have to relink Oracle to enable this.

What else was installed and configured on the Linux box.

What is a 'virtual partition' anyway? I'm going to be doing exactly this sort of work for a paper and presentation - the only really valid way to test is to use identical real hardware.

Finally where did you get the 30-40% figure from, I haven't yet seen any thorough published tests (rather than "we compared 2 unspecified systems and got a difference of about x% in unspecified tests" type posts).

All that said from the limited testing that I have done, both in vmware (which doesn't meet my criteria) and on a laptop which sort of does but isn't server hardware, your figures are broadly comparable to mine. I've also found that the windows system seems, surprisingly to me, to scale a little bit better (i.e I would expect the difference to be more marked at 90 users than 30).

You might also want to checkout swingbench from dominicgiles.com which is an oracle load tool written by an Oracle employee which allows a lot of good things including benchmarking your own code.

regards

--

Niall Litchfield
Oracle DBA
http://www.orawin.info
--

http://www.freelists.org/webpage/oracle-l Received on Fri May 19 2006 - 03:50:08 CDT