Oracle Performance and Engineering
Random notes on database features, code, practice, optimization and techniques.Tuesday, March 08, 2005
"Very Large SGA" and a few newsgroup posts
Today I worked on (ongoing) a document detailing how to configure large (> 1.7 GB but <>2.7GB) SGA with Oracle 9i or above on Redhat Advanced Server 2.1 or above. The document is by no means "complete" as of today. While most of the stuff is very well known, there are couple of interesting bugs to look at there.
As you surely know by now ;) - there was a big 'debate' going on in Tom's site. Whoa! Finally gotten the chance of reading all. Interesting in the way the thread has moved on from a proposition to disproof and then diversed into the ramblings on the "scientific methods" and all.
The technical stuff is interesting - if not useful. But it was great to read back the famous Einstein quote - "Intelligent people solve problems, geniuses prevent them". That, probably, is in essence the thread itself.
Someone surely is not very popular in Oracle newsgroup ;)
While it's definitely not true that Don Burleson is always wrong (his article on histograms in Oracle Magazine, for example, was very good and self-explanatory), but at the same time he's the guy who's more likely to state something generic without even a micro-level proof. Plus, I guess for a valid reason, Oracle Practitioners are kind of upset at sweeping statements without any validation.
Today's hilarious moment was definitely provided by Connor McDonald who proved index rebuild could (actually can never) reduce "clustering factor" only if...Isn't it funny?
As you surely know by now ;) - there was a big 'debate' going on in Tom's site. Whoa! Finally gotten the chance of reading all. Interesting in the way the thread has moved on from a proposition to disproof and then diversed into the ramblings on the "scientific methods" and all.
The technical stuff is interesting - if not useful. But it was great to read back the famous Einstein quote - "Intelligent people solve problems, geniuses prevent them". That, probably, is in essence the thread itself.
Someone surely is not very popular in Oracle newsgroup ;)
While it's definitely not true that Don Burleson is always wrong (his article on histograms in Oracle Magazine, for example, was very good and self-explanatory), but at the same time he's the guy who's more likely to state something generic without even a micro-level proof. Plus, I guess for a valid reason, Oracle Practitioners are kind of upset at sweeping statements without any validation.
Today's hilarious moment was definitely provided by Connor McDonald who proved index rebuild could (actually can never) reduce "clustering factor" only if...Isn't it funny?
Archives
January 2005 February 2005 March 2005 April 2005 May 2005 March 2006 April 2006 August 2006 September 2006 October 2006 January 2007 May 2008
