1.just exercise, if ok, I'll do actual production environment
Ok then I worry somewhat less for you ;-).
3.database runs about 2 years. So I think it need be processed.
This is hardly to be called a good reason for doing a reload of your database.
Any 'real' indication that you need to change something in your current database structure?
Like maybe large scatter factor for the large tables or fragmenation or other issues you encounter?
1.How to dump/load the data by binary?
proutil <dbname> -C dump ....
proutil <dbname> -C load....
2.How to dump/load the data by dd?
-use the datadictionary (obviously
)
-Make a program which dumps the df _user and sequences current values.
If these answers raise lots of questions to you, then I think it is best you hire a consultant or get some DBA training.
If you want to do a reload of your database because of problems you experience then you first need to think of what is the best way to solve those problems. Most reasons I know for people to do a reload is for performance reasons. If that is also the case for you, then it is better to redesign the database structure in such a way that you minimize these problems. You do this by:
-examining the database
-monintoring the database
-monitoring the OS
Furthermore you need to have knowledge of the behaviour of the data within the database:
-are there small but very active tables
-are there tables which fragment a lot
-are there some very big tables
- etc...
With this knowledge and measurements you can propose a new structure for your database.
Do some tests with this new structure and finally do the actual reload in the production environment.
What is the current strucuture of your database:
-do you already have storage area's?
-do you use type II storage area's?
-did you separate index from data?
-What version of Progress do you have?
-What OS are you using?
Regards,
Casper.