Okay, so I'm working with a new client, and remember that the admin side is not my strong suit, but...
Amongst many other questionable settings, they have a 40M initial exten setting for their primary data tablespace. Actually, they have 2 primary tablespaces: a large and a small one - the large one has an initial extent of 100M!!!!
Am I fully showing my ignorance or is this truly insane? I mean, it *is* a DSS database, but there are still some little code tables with less than 100 records that are in a tablespace with an initial extent size of 40M!!
What could possibly be the reasoning behind such a thing?
An initial extent of 40M is not unreasonable for a tablespace that will hold large objects. Code tables? Sounds like a waste to me. If they're worried about fragmentation, put the code tables in a seperate tablespace with more reasonable INITIAL and NEXT extents.
Last time when I was working with this database configured by people from Accenture I saw initial of 250MB and next of 500MB in an oltp database Waste of space (even thought the hd are cheap this is bad waste!)
In general, the next extent size should match the next 6 months data on that table.
The next extent size like 40MB / 250MB / 500MB is an estimation based on the past growth analysis.
Could it be, they intended to create the smaller tables with the physical attributes specified explicitly. (say, Create table small_table bla bla STORAGE (INITIAL small_size) )
Bookmarks