Data set placement recommendations are not mandatory. The
use of any or all of this layout is strictly optional. The recommended
layout provides a good foundation for many recent functional and system
management enhancements (such as sharing a master catalog, indirect
volume serial support, z/OS® UNIX, and Parallel Sysplex®). It can be used
to suggest how you might place data sets or as a set of objectives
to which you can evolve.
To help you decide whether to follow the recommendations, ask yourself
the following questions. Positive answers indicate that at least some
of the recommended actions would be beneficial to you.
- Will you require additional SYSRES logical extension volumes? Some SYSRES volume types, such as
the 3390-3, are not big enough to hold all the z/OS target libraries. Therefore, you might
have to move some data sets to SYSRES logical extension volumes.
Note: Several
data sets shipped with z/OS are
PDSEs. These data sets are CBC.SCCNCMP, CBC.SCLBDLL2, CEE.SCEEBIND,
CEE.SCEERUN2, SYS1.SHASLNKE, SYS1.SIEALNKE, SYS1.SIEAMIGE, and TCPIP.SEZALOAD.
One or more are most likely in your link list or will be added to
your link list as part of installation. If one or more are in your
link list and on your SYSRES volume, you should not share the SYSRES
volume beyond the scope of a sysplex because PDSE sharing between
sysplexes is not supported. If a PDSE is shared between sysplexes,
data could become corrupted.
- Do you plan to use indirect volume serial support? If so, the
recommended layout fits nicely with this support. See Using indirect catalog entries for more information.
- Do you plan to use ServerPac or SystemPac (dump-by-data-set format) for
future installations? ServerPac replaces at least your target and
distribution libraries, so using the layout recommended in this section
makes it easier to lay down the ServerPac or SystemPac (dump-by-data-set format) libraries.
- Do you plan to share master catalogs? The recommended layout
provides a good foundation if you do.
- Do you plan to move your volumes across systems? If so, the recommended
placement of user catalogs creates more-portable volumes and reduces
migration workload. Also, the master catalog alias resolution support
in DFSMSdfp allows
system symbols to be used for the user catalog name. See z/OS DFSMS Managing Catalogs for
details.
Based on these factors, you should determine which data sets to
place on each volume based on data set type, not based on element,
feature, or product. There are five types of data sets in the recommended
data set layout. Each type is placed on a separate (logical) volume.
The types of data sets and their volumes are:
- SMP/E global-shared data sets, on a volume shared by all
systems in the complex that need access to SMP/E global information.
See SMP/E global data sets for details.
- Target libraries (TLIBs) for product sets, on the following
volumes:
- TLIB volume 1 (TVOL1)
- TLIB volume 2 (TVOL2) through TLIB volume n (TVOLn)
- HFS or zFS target volume
- Licensed product set volume (for those licensed programs not installed
with the z/OS product set)
- Vendor product set target volume
- Subsystem product set target volume
See Target libraries (TLIBs) for details.
- Distribution libraries (DLIBs) for product sets, on the
following volumes:
- DLIB volumes for target volumes (which include TVOL1, TVOL2-n,
HFS, and zFS)
- DLIB volume for the licensed product sets
- DLIB volume for the vendor product sets
- DLIB volumes for the subsystem product sets
See Distribution libraries (DLIBs) for details.
- Image-related data sets, on the following volumes:
- Page data sets volume 1
- Page data sets volume 2 through n
- HFS or zFS customization volume
See Image-related data sets for details.
- Cluster-related data sets, on the following volumes:
- Master catalog volume (you can also choose to make this an image-related
volume)
- JES checkpoint volume
- JES spool volume
- Sysplex volume 1
- Sysplex volume 2
- Softcopy volumes
See Cluster-related data sets for details.
Many volumes on your system will contain data sets that are not
supplied by ServerPac or SystemPac (dump-by-data-set
format). Keeping such volumes separate from those that ServerPac or SystemPac (dump-by-data-set
format) will replace, or that you will replace when migrating the
new system to other system images, makes it easier to prevent overlaying
data sets that you want to keep. Volumes that contain non-ServerPac
or non-SystemPac data sets might include, but are not limited to,
the identified volumes in the data set descriptions, as well as volumes
for assorted data sets (dumps that were dynamically taken, logger
log streams, and so forth). Note that you can install your ServerPac
or SystemPac (dump-by-data-set)
order on volumes with existing data sets. See Preparing for installation for
the steps to follow when your target volumes contain data that you
want to preserve.
The rest of this topic contains details about the five types of
described data sets and volumes. As you read, keep the following
in mind:
- Volumes can be combined in order to conserve DASD. Combine volumes
based on like characteristics. For example, consolidating two SMS-managed
volumes with the same SMS constructs (Storage Class, Management Class
and in the same Storage Group) is more appropriate than consolidating
a non-SMS-managed volume and an SMS-managed volume.
- Although ServerPac and SystemPac (dump-by-data-set
format) considerations are mentioned specifically, this recommended
system layout is equally applicable to CBPDO users and will save time
when the new system is migrated to other images.