|
- CLUSTER
- CLUSTER defines or recatalogs
a cluster or cluster entry.
The CLUSTER keyword is followed by
the parameters specified for the cluster as a whole. These parameters
are enclosed in parentheses and, optionally, are followed by parameters
given separately for the DATA and INDEX components.
Abbreviation: CL
- NAME(entryname)
- Defines
the cluster's entryname or the name of each of its components. The
entryname used for the cluster as a whole is not propagated to the
cluster's components.
For SMS and non-SMS-managed clusters, the
component names must resolve to the same catalog as the data set's
cluster name.
You can define a separate entryname for the cluster,
its data component, and its index component. If no name is specified
for the data and index component, a name is generated. When the cluster,
data component, and index component are individually named, each can
be addressed. For information on system generated names, see z/OS DFSMS Using Data Sets.
When
defining a VSAM volume data set (VVDS), the entryname for the cluster
or the data component must be in the form SYS1.VVDS.Vvolser, where
volser is the volume serial number specified by the VOLUMES parameter.
The default primary and secondary allocation is 10 tracks. VVDSs cannot be defined in cylinder-managed space. For
information on defining a VVDS see z/OS DFSMS Managing Catalogs.
- CYLINDERS(primary[
secondary])
- KILOBYTES(primary[
secondary])|
- MEGABYTES(primary[
secondary])|
- RECORDS(primary[
secondary])|
- TRACKS(primary[
secondary])|
- The
amount of space in cylinders, kilobytes, megabytes, records, or tracks
allocated to the cluster from the volume's available space. A kilobyte
or megabyte allocation resolves to either tracks or cylinders; record
allocation resolves to tracks.
If the override indicator
in Data Class is set ON, the user specified amount of space in DEFINE
command will be overridden by the space values specified in data Class.
This enforces the installation standard of the system.
Exception: If
allocation resolves to tracks, the space is contiguous. See “Optimizing
Control Area Size” in z/OS DFSMS Using Data Sets for
more information.
Requests for space are directed to DADSM
and result in a format-1 or format-8 DSCB for
all entries.
If the cluster is not SMS-managed, you must use
the amount of space allocated, either through this parameter, or through
the DATACLASS, MODEL, or RECATALOG parameters. This parameter is optional
if the cluster is managed by SMS. If it is used, it overrides the
DATACLASS space specification. If it is not used, it can be modeled
or defaulted by SMS. If it cannot be determined, the DEFINE is unsuccessful.
If
you select KILOBYTES or MEGABYTES, the amount of space allocated is
the minimum number of tracks or cylinders required to contain the
specified number of kilobytes or megabytes. The
maximum space can be specified with unit of KILOBYTES or MEGABYTES
is 16,777,215. If the amount requested exceeds this value, you should
specify a larger allocation unit.
If you select RECORDS,
the amount of space allocated is the minimum number of tracks that
are required to contain the given number of records. The maximum number
of records is 16,777,215. If RECORDS is specified for a linear data
set, space is allocated with the number of control intervals equal
to the number of records.
The maximum TRACKS
or CYLINDERS value that can be specified on the DEFINE CLUSTERS command
is X'FFFFFF' or 16777215, because of the 3 byte space parameter
fields.
Recommendation: To maintain device independence,
do not use the TRACKS or CYLINDERS parameters. If you use them for
an SMS-managed data set, space is allocated on the volumes selected
by SMS in units equivalent to the device default geometry. If there
is an allocation failure due to lack of space, SMS retries allocation
with a reduced space quantity. However, any retry, including reduced
space quantity, is only attempted if Space Constraint Relief ⇒ Y is
specified. SMS also removes other limitations if the data class allows
space constraint relief.
Regardless of the allocation type,
the calculation of the CA (control area) size is based on the smaller
of the two allocation quantities (primary or secondary) in the DEFINE
command. A CA is never greater than a single cylinder, it might be
less (that is, some number of tracks), depending on the allocation
amount and type used. When tracks or records are used, the space allocation
unit (the CA size) can be adjusted to one cylinder. This adjustment
is made if the calculated CA size contains more tracks than exist
in a single cylinder of the device being used. The CA area size assigned
by VSAM is the smallest of: - One cylinder
- The primary space quantity
- The secondary space quantity
If the CA size assigned is not evenly divisible into either
the primary or secondary space quantity, VSAM increases that space
to a value evenly divisible by the CA size. If you are defining an
extended format data set, you should review "Defining an Extended
Format Key-Sequenced Data Set" in z/OS DFSMS Using Data Sets for
information about additional space requirements.
DEFINE RECORDS
allocates sufficient space to the specified number of records, but
factors unknown at define time (such as key compression or method
of loading records) can result in inefficient use of the space allocated.
This might prevent every data CA from being completely used, and you
might be unable to load the specified number of records without requiring
secondary allocation.
When multiple volumes are used for a
data set, these rules and conditions apply:
Secondary amounts can be allocated on all volumes available
to contain parts of the cluster regardless of the key ranges.
You
can specify the amount of space as a parameter of CLUSTER, as a parameter
of DATA, or as a parameter of both. When a key-sequenced cluster is
being defined, and the space is a parameter of: - CLUSTER, the amount is divided between the data and index components.
The division algorithm is a function of control interval size, record
size, device type, and other data set attributes.
If the division
results in an allocation for the data component that is not an integral
multiple of the required control area size, the data component's allocation
is rounded up to the next higher control area multiple. This rounding
can result in a larger total allocation for your cluster.
- DATA, the entire amount specified is allocated to the data component.
An additional amount of space, depending on control interval size,
record size, device type, and other data set attributes, is allocated
to the index component.
Note: If not specified, SMS
estimates the size of the index component to be 10% of the data component.
Ensure that enough space is present on the volume(s) to account for
this estimation.
To determine the exact amount of space allocated to
each component, list the cluster's catalog entry, using the LISTCAT
command.
The
primary and each secondary allocation must be able to be satisfied
in five DASD extents; otherwise, your DEFINE or data set extension
is unsuccessful. Starting z/OS® V1R13, if the primary or secondary space
allocation is 16777215 (X'FFFFFF') tracks or cylinders, the
value is decreased by 1 CA worth. This is because the AMS LISTCAT
command does not recognize all F's as a valid value and prints a zero
instead. - primary
- Allocates the initial amount of space to the cluster.
- secondary
- Allocates an amount of space each time the cluster extends, as
a secondary extent. You can use this secondary allocation to add space
for the data or index components of the cluster. A VSAM data set can
be expanded to 123 extents per volume. If this is a multi-volume VSAM
data set, then the VSAM component can be extended to a maximum of
255 extents combined over all volumes.
- VOLUMES(volser[
volser...])
- Specifies
the volumes on which a cluster's components are to have space. If
you do not use the MODEL parameter, or if the cluster is not SMS-managed,
VOLUMES must be used either as a parameter of CLUSTER, or as a parameter
of both DATA and INDEX.
VOLUMES can be specified or modeled for
a data set that is to be SMS-managed; know that the volumes specified
might not be used and result in an error. See z/OS DFSMSdfp Storage Administration for
information about SMS volume selection.
Note
that extending a data set to candidate volumes is different for non-SMS,
SMS and SMS guaranteed space VSAM data sets: - VSAM data sets allocated to volumes not under the control of SMS,
also known as non SMS VSAM data sets, have specific candidate volumes.
When a prime volume is full, the next specific candidate volume in
the list in the catalog entry is used. Should space be unavailable
on that volume, the next volume in the list is used. The process is
continued until either the allocation is satisfied or the candidate
list is exhausted. Thus, candidate volumes can be intermingled with
prime volumes as the list is not re-sorted after a volume is used.
Should sufficient space become available on a candidate volume that
was previously full, that candidate volume will be used.
- SMS-managed VSAM data sets that are not guaranteed space have
nonspecific candidate volumes in the catalog entry. When a prime volume
is full, SMS selects a volume that has sufficient space and causes
the VOLSER of that volume to be placed as the next prime volume in
the volume list in the catalog entry.
- SMS-managed VSAM data sets that are guaranteed space can have
specific candidate volumes with pre-allocated space. These volumes
are called candidates with space volumes. The pre-allocated space
for the candidate with space volumes is obtained when the VSAM data
set is defined. If the space cannot be pre-allocated, the define for
the VSAM data set fails. When a prime volume is full, the next candidate
with space volume in the list is used. Since the space on the volume
is already allocated, only the resetting of RBA values is necessary
to make use of the space and only the pre-allocated space is used
at that time. Specific or nonspecific candidate volume names (VOLSERs)
can be added to SMS-managed guaranteed space VSAM data sets after
the data sets is defined. These candidates follow the candidates with
space in the volume list in the catalog entry and are only used after
the guaranteed space candidates have all been used. These candidates
are used in the order of the volume list in the catalog entry regardless
of whether they have specific or nonspecific VOLSERs. The specific
VOLSERs are honored if there is space on the volume; if there is not
enough space, the extend fails. Nonspecific candidates then follow
the same process as SMS-managed VSAM data sets without guaranteed
space. Since SMS chooses a volume that has space, the extend will
always get the space unless the entire storage pool is full. Because
nonspecific candidates have a greater likelihood of succeeding for
an extend, it is preferable to have nonspecific volumes.
Letting SMS select the volume from the storage group
reduces the chances of allocation errors caused by insufficient space.
If the data set is SMS-managed with guaranteed space, SMS places the
primary quantity on all the volumes with sufficient space for later
extensions. If the SMS-managed data set does not have guaranteed
space or is a key range data set, primary space is allocated only
on the first volume. For SMS-managed VSAM data sets, the primary space
might be allocated on a different volume from the one you specified.
You
can let SMS choose the volumes for SMS-managed data sets by coding
an * for the volser with the VOLUMES parameter. If both user-specified
and SMS-specified volumes are requested, the user-specified volser
must be input first in the command syntax. The default is one volume.
For
SMS-managed and non-SMS-managed data sets, you can specify up to 59
volume serial numbers. If the combined number of volumes for a cluster
and its associated alternate indexes exceeds 59, unpredictable results
can occur.
If the data and index components are to reside on
different device types, you must specify VOLUMES as a parameter of
both DATA and INDEX. If more than one volume is listed with a single
VOLUMES parameter, the volumes must be of the same device type.
For
SMS-managed data sets, if you want the data and index components to
be on separate volumes for non-guaranteed space storage class requests,
code two different dummy names in the VOLUME parameter for each component.
If there are not enough volumes in the storage group to satisfy this
requirement, the allocation will fail.
If a guaranteed space
storage class is assigned to the data sets (cluster) and volume serial
numbers are used, space is allocated on all specified volumes if the
following conditions are met: - All defined volumes are in the same storage group.
- The storage group to which these volumes belong is in the list
of storage groups selected by the ACS routines for this allocation.
- The data set is not a key range data set.
The volume serial number is repeated in the list only
if the KEYRANGE parameter is used. You can use this to have more
than one key range on the same volume. Repetition is valid when duplicate
occurrences are used for the primary allocation of some key range.
If
a VVDS is being defined, only one volume can be specified and that
volume serial number must be reflected in the name indicated in the
NAME parameter.
If you define single volume zFS VSAM linear
data sets (LDSs) on a DEFINE RECATALOG command, the VOLUMES parameter
can have a special form referred to as indirect volume serial. The indirect volser must point to the original volume
that the zFS resides on. This results in the system dynamically
resolving the volume serial to the system residence (or its logical
extension) serial number. It can only be used with the RECATALOG parameter.
It can be used for both SMS and non-SMS managed zFS data sets on a
DEFINE RECATALOG of the data set. See the VOLUMES parameter in DEFINE NONVSAM Parameters for information on how to setup
an indirect volume serial.
When you clone a zFS, use the COPY
command with the PHYSINDYNAM (PIDY) parameter for the following reasons: - Using PHYSINDYNAM means you do not have to create a catalog entry,
but creates a catalog entry in the catalog for you. The catalog entry
is required by other methods.
- Using PHYSINDYNAM (PIDY) lets you use the same name for your original
and copied zFS.
For more information on how to clone a zFS, see Making
a copy of your system software (cloning) in z/OS Planning for Installation.
The
VOLUMES parameter interacts with other DEFINE CLUSTER parameters.
Ensure that the volume you give for the cluster is consistent with
the cluster's other attributes: - CYLINDERS, KILOBYTES, MEGABYTES, RECORDS, TRACKS: The volume
must contain enough unallocated space to satisfy the component's primary
space requirement.
- FILE: The volume information supplied with the DD statement pointed
to by FILE must be consistent with the information specified
for the cluster and its components.
Abbreviation: CYL, KB, MB, REC, TRK
Abbreviation: VOL
|