A dirty Datawarehouse

For a moment think of Your home refrigerator as a Database. You store stuff. You use that stuff. Process that stuff and then when fridge gets empty you refill it. (probably every week/month) . You also clean it up and wash it from time to time and keep it tidy.

Now imagine this fridge to be a giant Warehouse. An ill managed DataWarehouse. Imagine finding food thats stops looking like food (you know what I mean) . Real keepers of the warehouse long gone and mgmt kept a security guard at the gate who only knows to operate the on/off switch.

Something similar happened with me today after i was choosen to have a look at a 10 year old DW. The people who managed it are long gone. and noone knows whats where and the datafiles are a mess (so long to expect OFA ) . Which files are important and which are redundant !!! No info nothing.

If Tom Kyte were to look into this He would have a stroke. (May God give him long life )

Anyway First things first

1) At least bought the file structures in place.

2) MOnitoring for the schemas and objects in use

3) Patitioning based on month is a must.

4) This DB though production not in archivelog mode when i found it, which makes me wonder why !!!!

5) 100MB Redo log size !!! ( sounds not right !!! )

6) This DW is attached to Informatica/COGNOS

(I do NOT know anything about Informatica/COGNOS (as yet!) so if anyone can chip in some info about Oracle databases for Informatica/COGNOS that could be great anyway Google is everyones friend)

Its a good learning experience,

PS1: Are datawarehouses supposed to be dirty ?? what are your experiences

PS2 :On a personal note, I am trying to move to Delhi ( Presently I am in Pune) Lets see how soon that gets materialized.