SAS DI Jobs migration from Dev to Prod

[kakakapo]

Hi,

We can export SAS DI Jobs from Development Server and Import the .spk file in the Production Server. This is something we do as normal easy pracise.

Is there any other way we can achive this but not exporting and import functionality in DI Studio?

Problem is we have a very huge list of jobs to migrate and it takes time to do it in a relatively small time frame in production implementation.

Thanks,

kakakapo..

[LinusH]

What version are you using?

There is a batch interface that does the same thing as the wizards.

When using a wizard, there's an option to save/create the corresponding batch code.

What is that takes time? The manual export/import, or the processing time in the metadata server?

And then you need to consider how/which jobs to redeploy.

And, as discussed in earlier threads, redeployment can often come with a need to run addition scripts manually - if you are changing a physical table layout for instance.

[kakakapo]

We use DI 4.6 version and the number of jobs modified/newly developed would be migrated between Dev, SIT and Prod. Most of the times, we have to take individual jobs as .spk  export and import the same in the next environment. Sometimes, bigger packages have to be imported one by one at job/ table level and its painfull task.

When it comes to migration of jobs, some of them would be small and some would be bigger, time taken is more in the import and the export to be more precise. Re-deployment does not take much time to what ws observed.

Overall, I feel the time is taken when there is a communication made between Windows machine and the Server, When its already in the UNIX server, why bring them to Windows machine as a package and import it back to UNIX server using DI? Hope Batch does not involve windows.

So how do we do the same in batch interface? Would like to know how and if this could improve performance, would be bigger achievement.

[LinusH]

How - check out the documentation.
Performance - try it, performance is always hard to predict.