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

Home -> Community -> Mailing Lists -> Oracle-L -> Stress Test Ideas

Stress Test Ideas

From: Post, Ethan <Ethan.Post_at_ps.net>
Date: Wed, 18 May 2005 15:15:15 -0500
Message-ID: <E5D2567DDF0D03428A15664A78CA473C6B8607@pscdalpexch02.perotsystems.net>

I put together a little PL/SQL package which basically duplicates Tom Kyte's "runstats" script. I have added session_event and system_event info however, and it can handle up to 5 side by side compares at this point, it is easy to add more.

Another thing I am working on is a little "bench mark" script. Some ideas I have to "stress" the database are...

CREATE TABLE NOLOGGING | LOGGING DROP TABLE UPDATE TABLE (SINGLE COLUMN | MULTIPLE COLUMNS) UPDATE TABLE BULK (SINGLE COLUMN | MULTIPLE COLUMNS) DELETE TABLE INSERT TABLE (SQL) INSERT TABLE (PLSQL LOOP) HARD PARSE TEST - How fast can SQL statements hard parse.

REDO GENERATE TEST - How fast can I generate N MB of redo or how much redo can I generate in N minutes.

COMMIT TEST - How fast can I commit.

ROLLBACK TEST - How fast can I rollback a large transaction.

These are just some ideas I have, do you have any you can add that you think would be capable of providing some sort of decent benchmark.

I plan on combining the benchmark with the modified runstats output into a single script. I should be able to use the results to get a pretty good idea of what is fast and what is not so fast. The thing that is spurring this is that I may end of being responsible of upgrading (software application) a large customer database which I currently do not have access to and I want to know how fast these type of operations can be performed in relation to the test database I will be working with here.

--
http://www.freelists.org/webpage/oracle-l
Received on Wed May 18 2005 - 16:20:25 CDT

Original text of this message

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