TVOL1 is the first target library volume and the system residence
(IPL) volume. It contains many of the z/OS® target
libraries. Be sure to leave enough free space to allow for future
growth. TVOL1 allows you to IPL if one or more TVOL2 through TVOLn
volumes are temporarily not available.
TVOL1 contains some or all of the non-z/OS-UNIX target libraries for the z/OS product set. (See Product set for z/OS and closely-related products for
more information on the z/OS product
sets.) This does not include the licensed product set, which you install
separate from z/OS (on licensed
program volumes), or the subsystem product set (on the subsystem target
volumes).
The recommended types of TVOL1 data sets are as follows. Note that
the type of a data set is known within the Recommended System Layout
function in ServerPac and dump-by-data set SystemPac. The data set type can also be
found within the product's program directory.
- Load libraries.
- Change migration libraries. These libraries are used, or
might be used, during migration from one level of software to another.
- Help libraries.
- Panel libraries.
- Message libraries.
- Skeleton libraries.
- Table libraries.
- Fixed-block CLIST and EXEC libraries (if possible). If
you do not use fixed-block CLIST and EXEC data sets, then TVOL1 should
contain your variable-block CLISTs and EXECs and TVOL2 would contain
your fixed-block CLISTs and EXECs. (In other words, put the kind
of CLISTs and EXECs you use on TVOL1.)
- Data libraries. Some data libraries should go on TVOL2.
- SMP/E managed PARMLIB. This data set is the one pointed
to by the PARMLIB DDDEF, which will be used to store parmlib members
supplied by products you install. If you copy the SMP/E-managed PARMLIB
data set into your own system control data set, then the SMP/E-managed
parmlib should be placed on TVOL2-n. The placement of this PARMLIB
data set makes it easier to use concatenated PARMLIB support to reduce
migration workload.
- SMP/E managed PROCLIB. This data set is the one pointed
to by the PROCLIB DDDEF, which will be used to store JCL procedures
supplied by products you install. If you copy the SMP/E-managed PROCLIB
data set into your own procedure library, then the SMP/E-managed PROCLIB
data set should be placed on TVOL2-n. The placement of this PROCLIB
data set makes it easier to use concatenated PROCLIB support to reduce
migration workload.
Reasons to put these data sets on TVOL1 are:
- These data sets are critical to basic system function and are
required for recovery. Should some TVOLn volumes be lost or become
inaccessible, the system is less likely to fail if the most critical
data sets are all on a single volume. Also, in the event that all
TVOLn volumes are lost or become inaccessible, the ability to IPL,
log on, and use basic system functions after recovering TVOL1 can
speed the recovery of other volumes and greatly simplify the process
of restoring full function. (For example, restoring additional volumes
from backup tapes can be done in parallel, using cataloged tape data
set names.)
- Although it is technically possible to SMS-manage these data sets,
it is still recommended to not have these volumes under SMS management.
SMS volumes and data sets and volumes might not be accessible during
recovery situations (for example, ISPF edit). By extension, any data
set required for IPL is not recommended for SMS management.
- These data sets will be overlaid by a system replacement. By keeping
them together, you can more easily separate what will and will not
be overlaid.
- The SMP/E DDDEFed PARMLIB and PROCLIB are on TVOL1 so that TVOL1
can be IPLed in a minimum setup (without TVOL2 through TVOLn) and
still have the SMP/E-installed defaults available. The members stored
by SMP/E in the PARMLIB pointed to by the PARMLIB DDDEF, and those
stored in the PROCLIB pointed to by the PROCLIB DDDEF, are not sufficient
to IPL by themselves; a minimum set of system control parameters and
JCL procedures are required to IPL and use the system. You can either
concatenate your own PARMLIB and PROCLIB data sets to these data sets
(which is preferred) or add your own members to these data sets.