professional way of database and application study...???

Hi everybody,

I need one help from u all. Let me tell u first I am working for one big company in India since last 10 year. I am working as a Project Manager and Oracle DBA. I am certified Oracle 8i and Oracle 9i DBA worked on SCO-UNIX, Linux and NT We have lot of branches in India. Because of some reason we could not centralized database at mumbai and now we are going to do that in short time. Currently all the databases r distributed and all the branches sent only imp data by email and we insert those into main database.

Our Development (full in house package) which we are using since last 7/8 years is working fine (we have upgraded with d2k 6i) with all the branches including operation and accounts. Now same we have to centralize. So management had decided that lets study same package from professional people. So we have already appointed some XYZ company to audit our system. We thought we will learn something professional things from them. But I think we got something really funny report. Lets not go in to deep level. I want to talk only on 1 point .



We have Oracle 9i Database server with Data Guard Architecture (Stand alone in 8i).We have around 10GB Not so much. We are planning to go for Real Application Cluster.



Now important point is xyz company DBA and functional manager who studied our package and database suggested (just in 15 days they took to study we call them professional) that u should use 2 different physical databases one for operation and 2nd for only accounts. As per my experience and my knowledge concern I am totally against it. I feel there will be some major disadvantages; I hope u will also contribute some so I can tell management those…


Thanks in advance.

Disadvantage: -


1) Database performance will degrade because every time application has to give two call to two separate databases.

2) Entire application will have to change, which is not so easy.

3) License will have to purchase for two servers O/S Linux as well as for Oracle databases.

4) Lots of Coding has to make for transferring data from 1 database to 2nd database, which will really create problems.

5) Operation data and Accounts data which is so important for any company will wont be flow online (Any program or report has to be based on 2 different server)

6) Changes made into operation database has to be made again into accounts, vise-versa

7) Both server required backup server (Data Guard) or if we go for RAC environment then again both server required more server which will have to maintain properly.

8) Disaster and Backup strategy have to implement properly for both server.

9) If any one operation or accounts (Production) server failed then entire application will come to halt till problem get solved.

Smash,
I'm trying to understand what it is your saying. Correct me if I'm wrong, but you've had several databases now you're consolidating them all together. You've had "professionals" come and audit your systems and they have said that it would be benifical to have two different databases but having the same data?

1. You are correct 10gig is very small. My first question would be how many users will hitting this database and can your present stand alone server handle the load?

2. If you feel that your present hardware situation can handle the present load plus future growth then I would submit documentation to counter what the "professionals" have offered.

3. Read this, it's one of my all time favorites concerning RAC.

For which reason recommand they to separate account and operation?
What should be the advantage?
:confused:

Quote:

---

Originally posted by mike9
For which reason recommand they to separate account and operation?
What should be the advantage?
:confused:

---

Indeed, they ought have listed costs and benefits of each approach -- if not then their credibility would be very poor, I think.

Thanks for all the reply.

Regarding oracledoc query .....We have only 1 main database which has around 20 schemas ....all schemas are linked to each other...like masters....reference keys.....Each schema is depend on each other..

This is just example...(not real one)
we have master schema....(emp_master,dept_master)
operation schema .....
transaction schema ..which hold accounts data....

Now what they are saying split all schemas ...keep only 1 accounts schema separate on different database,I know this is almost impossible. Because every time accounts has to refer operation database for masters and for each and every transaction for details.

We are using Xeon server with dual processor. 36 GB x 5 .RAM 4 GB Number of users will be hitting database will be around 350 to 400. We may go for good server that is not a constraint.

Mike …they have not studied properly I am sure. …there will be more disadvantages then benefits …

Please comment oracle Guru ..I am waiting for your help…

Thanks ....

You are right......credibility would be very poor....or study is not proper....how can u study full system which is working since last 8 years in 15 days time.... gr8........

I agree to it...

Indeed, they ought have listed costs and benefits of each approach -- if not then their credibility would be very poor, I think.[B]

Most of your arguments are pretty weak. Some can be addressed by technology already included in the base software, some in EE Options. In addition, licensing should not be an issue because you are going from n+x servers to n. Your license costs are going to be the same or less anyway.

I think your point that brings the most weight is:
For example:

Quote:

---

Originally posted by smash

9) If any one operation or accounts (Production) server failed then entire application will come to halt till problem get solved.

---

If this is truely the case, you would be better putting both apps in the same instance. If, however, you can still run one application with stale data from the other, this argument too is not valie.