Bypassing a package build?

[MagarG]

I was seeing if anyone has done this before or if it is possible. After talking to JDE and some others, they say it may be better just to build and deploy instead of trying to manually do it to make sure nothing is missed. Say, we have PY all ready to go. We have build the full client and server packages and tested. PD is going to be identical to PY except for the data of course and OMW has transferred all the objects to PD. We want to build FULL packages for PD. Can I somehow, copy the PY client and server packages without going through the build process.

I was talking to JDE and it appears when you copy a definition or build, the directories behind the scenes are not created or copied on the deployment server. JDE said if the environments are identical, you can manually copy the path code library (AS400), the IFS folders (AS400), and the inf files (package_inf folder DEP server), then change the .inf file to reflect the correct path code. On the DEP server, for example, to get the main package files:

copy \B7333\PY7333\PACKAGE\PYFULL1 to \B7333\PD7333\PACKAGE\PDFULL1

There, however, are setup.inf files located on the DEP server under each package (located in \include, \make, \res, \source, \work) which have paths in them but haven't been modified since the install of Xe. They are also not text files. Would I need these files or were they just for the initial installation?

tia, Grant

AS400 V4R5 SP15 Xe
JAS SP15
WebSphere 3.5.3 on AS400




AS400 V4R5 CO-NT Xe SP15 Oneoffs 002/007/010 JD9533 ESU
Websphere 3.5.3 on AS400 JAS SP15

 

[Carl_Fisher]

Yes this is possible - it is what we did when we went live with Xe.

The only thing you need to do is create a PD (in Work With Packages, this
creates the Database entries to allow you to install the package.

I would copy everything from the Package\PDFULL1 folder.

We even created our PD7333 folders on the deployment server and enterprise
server by copying the PY7333, to ensure that what was live was what was
tested.

There may be a few other fiddles you need to get it to work - I think you
have to copy the Object Librarian and Versions records, as these are
environment specific, as we are an Oracle site we use SQL for that.

We were advised by JDE all the way - in fact it was one of their technical
people who came on site and helped us with the work, so they should know all
the tricks, if you can get the right person.



OW733.3 Xe SP 14.2
Enterprise Server - Intel NT + Oracle 8.0.6
Client - Citrix TSE + 4 NT PC's for development

 

[Spike]

There's a document out on the Knowledge Garden that will tell you how to do this. It lists all the steps required to do this. I don't remember the document number, otherwise I would have included it here.



Ryan Allen

OW Xe B7333 SP14.2 / World A7.3 cum 12
ES - AS/400 V4R4
DS - Win2000
Fat - Win200
TSE/Citrix - Win2000

 

[sperkins]

This sounds like a bit of techie fun to prove a point, but can someone
explain to me why you would actually want to do this? What are the
advantages? What benefit is there to the business of such a manoeuvre and,
more importantly, what are the downsides to this?

I await enlightenment. Cheers.

OW Xe, Sandpit Windows2000, SQL2000, Compaq ML370 x 2
SP15
Update1

Sid Perkins
Associate Director &
JDE Technical Project Manager
CB Hillier Parker Limited
Tel: 020 7882 8364
Fax: 020 7882 8720
Mobile: 0797 953 0506
www.cbhillierparker.com

 

[MagarG]

Re: RE: Bypassing a package build?

We are setting up another path code and to save time building the initial package (the time it takes to do a full client and server package build plus deployment of the server package) we are trying this. We just got done manually copying a full PY package to PD and it's working. The longest part of this manual process is copying the deployment server package to another path and copying the library on the AS400 to another path. So far it has saved us over 10 hours of package building time because we did not have to build the client and server. We are now trying to do a partial package.

There is a document on the KG that eludes to this from the GATS called "Creating Initial Installation Packages for Custom Path Codes/Environments Without Running a Package Build".

AS400 V4R5 SP15 Xe
JAS SP15
WebSphere 3.5.3 on AS400

This sounds like a bit of techie fun to prove a point, but can someone
explain to me why you would actually want to do this? What are the
advantages? What benefit is there to the business of such a manoeuvre and,
more importantly, what are the downsides to this?

I await enlightenment. Cheers.

OW Xe, Sandpit Windows2000, SQL2000, Compaq ML370 x 2
SP15
Update1

Sid Perkins
Associate Director &
JDE Technical Project Manager
CB Hillier Parker Limited
Tel: 020 7882 8364
Fax: 020 7882 8720
Mobile: 0797 953 0506
www.cbhillierparker.com





--------------------------


AS400 V4R5 CO-NT Xe SP15 Oneoffs 002/007/010 JD9533 ESU
Websphere 3.5.3 on AS400 JAS SP15

 

[MikeGibson]

I don't think you can get out of doing the actual package build. When you do the copy from PY to PD you copy everything but the Package directory. This insures that all your source code is the same. You also need to copy your central objects from PY to PD and copy versions from PY to PD and update this table so it points to PD instead of PY after the copy. After doing these steps PD source and objects should be the same as PY. Now you will have to create your PD full package so that it can build and be able to deploy it. If you don't build a package how will you deploy it? There maybe someway of fooling it by renaming a full package under PY to a full package that was previously built for PD when you copy the source.... sounds kind of risky though.

We did this as we went live on XE to. To answer a question about why, it saves a lot of time. If you prototype and test in one environment, you need a fast way of getting your production objects updated with all the changes.


XE SP15, Oracle 8.1.6 on W2K,Citrix/W2K, NT 4.0