Version Control in DataStage

[WoMaWil]

Hi Sam,

there are several things you should analyse to solve your Problem. As you already know there are two ways to migrate Jobs from on DataStage-Project on a server to an other project on maybe another server. One is Releasing and the other is Export/Import. Both with its advantages and disadvantages, which can nevertheless be solved. Finaly it up to you, your team and your Managers to find the solution which fits your needs.

In DataStage you have 4 different roles.

(1) none, this people can do nothing within DataStage
(2) Operators, this people can start and schedule job, view logfiles and other details. Their Tool is only the director.
(3) Developers. They can do all of (2) and beside this develop jobs. Their additional tools are the Manager and the Designer
(4) Administrator, (until version 4.x this was root, after that it is the user dsadm) they can do all of (3) and adminstrative things like adding projects and so on.

If you migrate via Export/Import the Jobs on the Production Server can be seen without any problems from Developers on this machine. Because the Operators have only access to Director, they cannot see the Job-design. They only see the logs and other director details.

If you migrate via Releasing Jobs the same is true for Operators and Developers, but the Jobs can't be changed on the target machine. You can change only on the developer side, which makes the process more secure.

You are right, the migration via source safe is a bit complicated. Therefore for example we have written a little Dos/windows-commandline-tool for to export, import and compile jobs. We are with this tool able to start our versioning from within our versioning tool. (In our case PVCS) and this makes versioning very easy.

Our company does sell this tool. If you need additional details please contact me Wolfgang.Huerter@gerling.de.

Wolfgang