on which machine is a package built

If I am working on a fat client on my desktop, and submit a package build, where is the build processing running? is it my desktop or enterprise server or deployment server?
Can I login fat client in PY env, and submit a DV package?

We are runing E811SP1, 8.96D1, SQL2000 all on windows

Hi,

UPDATE PACKAGE BUILD
1.
a) For deployment of ESU's, login to DEPLOYMENT server using DEP81x (deployment server environment)
b) For custom objects, login to FAT client it was created or use GET to bring down the specs to the fat client if it was not customized on that workstation and login to the deployment server environment. We do this to assemble to objects only and then change to deployment server to build and deploy the package.

2. Now assemble, build and deploy the package build.

As for us, we assemble the package on the fat client and use the deployment server to build and deploy the package. We do use the deployment server as it has a higher specs than our regular workstation and it's under utilized as i see it. So might as well use it for its purpose.

Package build submitted via the fat client will use the resources of the fat, so if you use the deployment to submit the build it uses the deployment servers resources. Deploying the package from the deployment server uses also a portion of the enterprise resources.

Don't mix and match the different environments, you'll never know what you gonna get...

ciao

[ QUOTE ]

Package build submitted via the fat client will use the resources of the fat, so if you use the deployment to submit the build it uses the deployment servers resources. Deploying the package from the deployment server uses also a portion of the enterprise resources.


[/ QUOTE ]

This is totally misleading.

The package build process will always convert Central Objects into Replicated Objects. Central Objects are stored at the Database Level and a directory on the Deployment Server. The resulting "package" is stored on the Deployment Server.

A package build will build objects using the CENTRAL objects every time - so even if you do not have the latest package in your client, or even if you don't have the pathcode installed (for example, you are building a PY package from your DV client) - it will still use the correct specifications to build. The only time this is an issue is if you try to build on the deployment server logged into the JDEPLAN environment (don't do that).

There is no issue using a development workstation to build a package (as long as the compiler is present). Unless you have some weird OCM mappings, you'll get the same results every time. Wherever the build is being processed is where the CPU is being used. You can also use a Terminal Server, providing it has been set up as a Development Terminal Server.

The enterprise server will only be used to convert the TAM files from the client (PAKTOTAM) and to compile the business functions.

Hope that explains the package build process.

very informative answer, thanks.

Further the discussion, it seems OK for me to deploy package from any environment on any fat client, I don't need to RDP to deployment server and login to DEP811 environment to deploy package.
I don't mean to run against the best practice, just curious on how things working behind the scene. Why we need DEP811 if I can do all the tasks in other env?

Thanks