When there is a need to change the encyclopedia repository from the Gen z/OS Host Encyclopedia to the Client Server Encyclopedia (CSE) or just to offload models from a Host Encyclopedia to a CSE Encyclopedia, here are some tips to consider.
Release: Any
The Gen Host encyclopedia software has built-in capabilities to move a model to the CSE as a child model but not as a "normal" parent model. A parent model has the characteristics that it is primary to the encyclopedia in which it resides. It is not expected to be returned to the Host encyclopedia so there is no limitation on the process and functions that can be performed on the CSE. A child model is limited in its functionality e.g. Version Control. So when downloading models from the Host Encyclopedia, there are steps to be followed to ensure that the models are parent models.
There are two major reasons models loaded to the CSE from the Host must be parents. The destination model in version control must be a parent. An error will occur if the destination model is a child model. The second is if the models remain child models and developers do updates, extraneous objects and associations will be retained in deleted status needed to upload models back to the parent (Host). This situation will cause problems with processing other functions of the tool.
The download process from the Host Encyclopedia to the CSE will result in a child model on the CSE. There is no automatic process such as the UNLOAD / LOAD function in DB2 for moving all models at one time from the Host encyclopedia data tables to a non-DB2 database on the CSE, e.g., Oracle. Models must be downloaded individually from the Host encyclopedia to the CSE.
All subsets and models that will be moved to the CSE must be checked-in to the Host Encyclopedia before downloading to the CSE. A full check-in of all models/subsets is a prerequisite action for transferring a model permanently to the CSE.
The following steps will move a model from the Host to the CSE with ancestry retained:
Download Model COMMAND ===>Type the Model name or request Prompt for list selection.
Model name . . . . . . GEN SAMPLE MODEL________________________ + Software version . . . 9.1.A5 Transaction file name. IEF.TRAN________________________________The extract option will create a LOAD transaction file to be uploaded to the CSE. Again, this method will retain ancestry. Adoptions are not necessary before migrations.
Upload option . . . . / No . Yes Extract option . . . . . No / Yes Child ency id . . . 1________
Child model name . . GEN SAMPLE MODEL CSE____________________
(blank if same as parent)
Codepage . . . . . . . 1252_ Readonly option. . . . / No . Yes Execution mode . . . . / Online
. Batch
F1=Help F3=Exit F4=Prompt F12=Cancel
upload -u <userid> -t load -d <directory>
The following steps will move a model from the Host to the CSE without retaining ancestry:
upload -u <userid> -t update -d <directory>
If applications will continue to be generated to run on the Host Mainframe, e.g., CICS, IMS, then the Gen z/OS Implementation Tolset (IT) is used to build them. Once the models are on the CSE and have been correctly prepared for use, changes to the model can occur on the CSE or the workstation toolset. The model or subset can be downloaded to the workstation toolset for further object changes before remote generation. MVS bind properties to set collection Id, version, and owner can be set using the MVS Bind Options for Triggers or the Business System in the Construction Client. The dynamic link of procedures, action blocks and screens is set using the Environment Parameters window in the Construction Client.
A remote file can be generated from the Construction Client or the workstation toolset by setting the generation default parameters for DB2 z/OS. A remote file is created which needs to be transferred to the Host in ASCII (TEXT) format for use with the z/OS IT. The z/OS IT will process and install applications as specified in the target, e.g., CICS, IMS, TSO. When using the z/OS IT, the same Host Construction datasets that were used when generating applications on the Host encyclopedia can be used as location datasets in the z/OS IT.
If applications will be moved to a new OS platform and Programming Language/DBMS then the corresponding Gen Implementation Tolset (IT) for that platform is used to build them. Once the models are on the CSE and have been correctly prepared for use, changes to the model can occur on the CSE or the workstation toolset. The model or subset can be downloaded to the workstation toolset for further object changes before remote generation.
A remote file can be generated from the Construction Client or the workstation toolset. A remote file is created which needs to be transferred to the target OS platform in ASCII (TEXT) format for use with the IT. The IT will process and install applications as specified in the target.
NOTE: If the target platform is Java then a Gen Windows IT is required.
Relevant doc. references:
Gen™ 8.6 > Encyclopedia > Host Encyclopedia > Using the Host Encyclopedia > Manage Models in the Host Encyclopedia > Download Model
Gen™ 8.6 > Encyclopedia > Client Server Encyclopedia > Use the Client Server Encyclopedia
Gen™ 8.6 > Encyclopedia > Client Server Encyclopedia > Use the Client Server Encyclopedia > Encyclopedia Communication
Gen™ 8.6 Implementation Toolset
Related KB article: Gen Encyclopedias and importance of unique Encyclopedia IDs
NOTE: The default HE Encyclopedia ID will be 1 but in practice, it may be different (Gen™ 8.6 > Installing > Install Gen on z/OS > Configuring Host Encyclopedia and Host Construction > Configuration Variables)