I am working to upgrade the zPDT environment to the latest ADCD release of V2.2.  I currently have the C2 volumes and looking to upgrade to the D2 volumes.

The way I built out the system was to keep all our customizations separate from the system stuff.  Having a concat list, etc with my custom libraries/parms at the top.  Basically all I need to do to upgrade is replace the system volumes with the new ones.  Migrate over the RACF dataset and copy in my IPLPARM member.

The issue that I came across was that none of my datasets on any of my volumes were cataloged.  I realized what was happening.  With the new system supplied packs, there was a new master catalog so obviously none of my volumes were in there.  I didn't create any user catalogs since this is really a low activity system.

So my question is, what type of catalog strategy should I use to ease this upgrade path?  My initial thoughts with some research is to lock down the master catalog and create user catalog(s)/aliases for all my new dataset HLQs.  Then link that stuff to the new master catalog on each upgrade.  I still have some SYS1 type of datasets that were customized that I'll have to uncatalog the supplied ones and catalog my new ones.

Looking for some input on this type of strategy or a better strategy.

Thanks.

John, all your assumptions are correct. You actually need to define a user catalog and define an alias to your datasets HLQ.

Our recommendation is the following:

1. Before start using ZDT and ADCD z/OS package, spend some time on a planning for you purposes of usage of this environment. Consider how much space you will need, dataset names pattern, aplication requirements, subsystem configuration/definition, etc. 

2. Allocate a new private disk to be used as your work space.

    This process will require the following steps, just in case you need to know:

     - use the command awsalckd to format a Linux file

     - include this new file in the devmap file

     - stop/start the emulator (awsstop/awsstart) to load the new device

     - After IPL, use ICKDSF utility to INIT the disk to z/OS

     - vary this new disk online and start using it.

3. Allocate in this new disk a PDS or PDSE to put your JCLs for future use.

4. Create in this new disk a user catalog (normal ICF catalog) and define an ALIAS according to your plan. Even small ICF catalogs allows thousands of datasets on it, so a space like 5 or 10 CYLs should be more than enough.

5. If you need SMS-managed datasets for any reason, remember to add this new disk in the SMS constructs and in the ACS routines. 

When you update your ADCD package, you don't need to copy everything again. Just keep the Linux file with your private disk and replace only the ADCD disks.

After you IPL, just re-execute the JCL to CONNECT your user cat to the new master catalog.

Additionally, if you need to update any PARMLIB member or z/OS configuration dataset, it's recommended to create your own PARMLIB, PROCLIB, ISPF datasets and others on your own disk and concatenate them in the z/OS datasets. If you need additional information about how to do it, please learm more in the following books or Knowledge center:

- z/OS Init and Tuning Guide

- z/OS Init and Tuning Reference

- JES2 Init and Tuning Reference / Guide

Thanks for the information.

When I started working with the zPDT, I had zero MVS mainframe experience, so you learn as you go along.

With this mini-up upgrade, I'm fixing things such as the catalogs in preparation of the next upgrade.

I do have a lot of customizations built into my instance a lot of it is to test some things before we implement them on our prod LPAR.

A couple things I have noticed, in addition to my lack of knowledge when setting zPDT up for the first time, for this upgrade from Z22C to Z22D:

---The USER libraries now include the system version as part of the dataset name.  So some overridden parms and procs needed to be changed to account for this

---Some datasets seem as if they need to be in the master catalog or they won't be found during the IPL.  Before I cleaned up my usercats everything was in the master catalog so I never noticed this.

---SMS storage groups need some work.  I'm having an issue when I can't delete, uncatalog, catalog datasets in my storage groups.  Need to figure that out yet.

---Some other minor tweaks.

When I get a all completed and resolved I'm going to document my process and post it here so it may be helpful to others in the future.

Thanks

Great idea sharing your experience. Appreciate.