Any tips, tricks, or words of wisdom out there regarding the tuning of Parallel Queries? Besides just setting parallel_automatic_tuning = true and hoping for the best :).
- Chris
Printable View
Any tips, tricks, or words of wisdom out there regarding the tuning of Parallel Queries? Besides just setting parallel_automatic_tuning = true and hoping for the best :).
- Chris
SQL hint like this: /*+PARALLEL (TABLE_NAME, 2) */ ...
For example:
SELECT /*+PARALLEL (TABLE_NAME, 2) */ TABLE_FIELD1 FROM TABLE_NAME
There's a good discussion in Oracle Performance Tuning Tips & Techniques by Richard Niemiec if somebody in your company has it. (Chapter 11)
Thanks Jeff.
Ironically enough - that is one of the 2 Oracle books that I *do* own. I'll have to check it out when I get home. To tell you the truth, though, I only got but so far in that book. His section on SQL Tuning completely turned me off. His ideas and methodologies on that subject were so old-school and antiquated that I couldn't read more than a few pages at a time without my blood pressure rising to undesirable levels :). I'll give him the benefit of the doubt on this topic, however, since it's coming from you ;).
Thanks for the tip,
- Chris
I agree. You have to know enough to weed out the O7 stuff from the newer stuff. That's what happens when you keep "upgrading" your book I guess...Quote:
Originally posted by chrisrlong
Thanks Jeff.
Ironically enough - that is one of the 2 Oracle books that I *do* own. I'll have to check it out when I get home. To tell you the truth, though, I only got but so far in that book. His section on SQL Tuning completely turned me off. His ideas and methodologies on that subject were so old-school and antiquated that I couldn't read more than a few pages at a time without my blood pressure rising to undesirable levels :). I'll give him the benefit of the doubt on this topic, however, since it's coming from you ;).
Thanks for the tip,
- Chris
Hearing this from Chris it comes like a music to my ears ;). Rich Niemiec's book was the only book about Oracle so far that realy made me *very* upset when I read it last summer. There is so many missconceptions, doubtfull "facts" and rull of thumbs and so many wrong conclusions that I started to make a note of each of the things that I thought were wrong or doubtfull. At the end it came out as a 12 page long MS Word document (together with my explanations of why I thing they are wrong) and I've sent it to Rich. I got an answer that he is bussy and at the moment but that he will study my remarks carefully later and let me know about his opinion on my comments. I never got that second mail, but I've talk to Rich at last year's EOUG meeting in Madrid. He said that he aggrees with some of my observation, yet disaggree strongly with some (most?) of them, but he never took the time to send me any serious answer to my remarks on his book.