This topic describes the tasks that must be performed to use DFSMShsm,
a component of DFSMS that
manages your storage in a hierarchical manner. It contains examples
of the commands needed to perform the tasks, descriptions of the processing
that DFSMShsm performs,
and examples of the results of that processing.
Introduction
Welcome to DFSMShsm explains how to use this document and describes
the functions of the DFSMShsm program, highlighting the following
information:
- The tasks that you perform to make DFSMShsm do useful work.
- An overview of how DFSMShsm moves data.
- How to use this document. You are introduced to the example system,
which is referred to throughout the rest of this document.
Space management
The following topics introduce the tasks necessary to control space
management functions. Space management is the DFSMShsm program function that
you use to ensure that your customers have space available on DASD
volumes to allocate new data sets or to extend old ones. You can make
the space available by deleting data sets that have outlived their
usefulness, removing unused allocated space from data sets, moving
low-activity data sets from level 0 volumes to other DASD or tape
volumes, and returning the moved data sets to the level 0 devices
when the data sets are needed.
Functions of space management:
- Automatically delete, remove unused space from, and move data
sets on DFSMShsm-managed volumes to provide a specified amount of
free space on each volume. The space management functions begin at
a specified time of day and only on a specific day of a cycle. This
operation is called automatic primary space management. The process of removing unused space from the data sets is
called space reduction. Moving the data sets
(as opposed to deleting or removing unused space from them) is called migration.
- Reconnect unchanged data sets that are recalled from ML2 tapes
to their original ML2 tapes with the fast subsequent migration function.
- Ensure that a specified amount of space is available on DFSMShsm-managed
volumes. If DFSMShsm finds any volume without the specified amount of space, it performs
migration from all such volumes (known as event-driven
migration) following a space check.
- Automatically clean up migration volumes and the migration control
data set (MCDS). This function and the following one are known as automatic secondary space management.
- Automatically, or by command, migrate data sets from level 1 migration
volumes to level 2 migration volumes.
- Automatically recall needed migrated data sets to level 0 volumes.
- Recall migrated data sets to level 0 volumes by command.
- Delete eligible data sets on a non-SMS-managed volume by command.
- Delete eligible non-SMS migrated data sets on a migration volume
by command.
- Migrate individual data sets or eligible data sets on a volume
by command.
Availability management
The following topics introduce the tasks necessary to control availability
management functions. Availability management is the function of the DFSMShsm program
that you use to ensure that your customers can retrieve usable copies
of their data sets should their online copies become lost or damaged.
You can make the copies available by making daily incremental backup
copies of changed data sets, making periodic dump copies of the DFSMShsm-managed
and ML1 volumes, making aggregate backup copies of data sets that
your operation will need if your installation is damaged, and making
fast replication backup versions for sets of storage groups.
Functions of availability management:
- Automatically make backup copies of individual changed data sets
on DFSMShsm-managed volumes. This is known as incremental backup.
You can specify how frequently to back up data sets on a data set
basis for SMS-managed data sets and on a system-wide basis for non-SMS-managed
volumes. You can specify on a system-wide basis how often backup runs.
- Automatically make dump copies of DFSMShsm-managed level 0 volumes
and ML1 volumes on a specified schedule for day and time. You can
dump different groups of volumes on different days with different
periods for the number of days between dumps.
- Allow your customers to issue commands to recover their own data
sets.
- Allow your customer to issue the HBACKDS command from a batch
environment. See Inline backup.
- By command, back up user data sets.
- By command, back up data sets of an application to tape so they
can be taken to another computer site for recovery.
- By command, recover user data sets or data sets of an application
to their original system environment at another computer site.
- By command, restore a volume from a dump copy and update the restored
volume from later incremental backup versions.
- By command, recover a specific data set from either a dump copy
or an incremental backup version.
- By command, create a fast replication backup of a copy pool.
- Automatically or by command, create a dump copy of a fast replication
DASD backup version
- By command, recover volumes that have a fast replication backup
version.
- By command, recover data sets from fast replication backup versions
Backup function
Backup is the process of copying a data set from a level 0 or an
ML1 volume to daily backup volume. This copy is called a backup version.
The purpose of backup is to have copies of data sets in case something
happens to the original data sets. The difference between dump and
backup is that the dump function backs up the entire allocated space
on a volume, whereas the DFSMShsm backup function backs up individual
data sets.
DFSMShsm can
create backup versions of data sets either automatically or by command. DFSMShsm automatically
creates backup versions of data sets on specified days beginning at
a specified time of day. The data sets must meet eligibility criteria
and must be on DFSMShsm-managed volumes that have been designated
for automatic backup.
Fast replication function
Fast replication uses volume-level fast replication to create backup
versions for sets of storage groups. You can define a set of storage
groups through the SMS copy pool construct. The DFSMShsm FRBACKUP command creates
a fast replication backup version for each volume in every storage
group defined within a copy pool. The fast replication backup versions
can be dumped to tape. The dump copies can be created by using the
FRBACKUP command or with Automatic Dump processing. You can use FRBACKUP
to dump an existing fast replication backup version, or when no backup
version exists, create a fast replication backup version and dump
it to tape. In all cases, the dump copies are associated with the
original source DASD volumes in the copy pool, but not with the target
DASD volumes that are actually dumped. The dump time stamps reflect
when the fast replication backup version is made, not the actual time
of the dumps. Recovery from a fast replication backup version can
be performed at the copy pool level from a disk copy, at the individual
volume level from a disk or tape copy, or at the data set level from
a disk or tape copy. This Fast Replication function enables the backup
and recovery of a large set of volumes to occur within a small time
frame.
The term fast replication refers to the FlashCopy® function supported by IBM® System Storage® disk and the SnapShot function supported by IBM RAMAC Virtual Array (RVA) disk.
Dump function
Dump is the process of copying all data from a DASD volume to dump
tape volumes. Full-volume dump is an extension of DFSMShsm’s
availability management that invokes DFSMSdss through the DFSMSdss application interface.
Full-volume dump backs up the entire allocated space of DFSMShsm-managed
DASD volumes and ML1 volumes either automatically or by command. Non-DFSMShsm-managed
volumes are dumped only by command. The purpose of the full-volume
dump is to expedite the recovery process when an entire volume is
lost or damaged and to supplement the incremental recovery process.
The full-volume-dump process can be one volume in and one dump
copy out. It can also be one volume in and multiple dump copies out.
Each dump copy is a complete image of the dumped volume, not just
an incremental backup of selected data sets. Each successive full-volume
dump of a volume, regardless of the number of dump copies, is a generation.
Each dump copy in a generation is associated with a different dump
class, which specifies how the dump copy is to be managed. DFSMShsm allows
from one to five dump copies to be made concurrently for any one full-volume
dump. DFSMSdss discontinues
writing to dump copies on which errors occur. DFSMSdss creates the multiple copies
and continues its full-volume-dump process as long as one output copy
is good. DFSMShsm discards
the contents of only the bad copies and issues an appropriate message.
If all copies fail, the full-volume dump is failed.
DFSMShsm keeps
100 generations of dump copies for any given volume unless all the
copies in a generation expire. When all dump copies in any generation
for a volume reach their expiration dates, DFSMShsm deletes that generation
from its records. When the 100 generations for a particular volume
have been reached and the next full-volume dump for that volume is
performed, the control records for the oldest generation are discarded,
regardless of the retention periods of the individual copies.
Types of backup volumes
You can define the following types of tape or DASD backup volumes
to DFSMShsm:
Recovery and restore of data sets
Recovery and restore are processes that are requested only by command,
not automatically, for backed up data sets. Recovery is the process
of retrieving a full-volume dump or a backup version of a data set
or a volume. Restore signifies that DFSMSdss is used to retrieve dumped data from
dump volumes. You can use restore or recover processing to:
- Recover a data set that has been lost or damaged.
- Recover an earlier version of the data set without deleting the
current version.
- Restore a volume from a full-volume dump and update the volume
from later incremental backup versions.
- Restore a data set from a dump copy.
- Restore a volume from a full-volume dump.
- Recover a volume from DFSMShsm backup versions.
- Recover a pool of volumes from the fast replication backup versions
on DASD.
- Recover a single volume from the fast replication backup versions
on DASD or tape.
- Recover a data set from the fast replication backup versions on
DASD or tape.
DFSMShsm volume
recovery can use incremental backups or full-volume dumps or both.
One DFSMShsm RECOVER
command can be used to request both a volume restore and an incremental
volume recovery.
DFSMShsm can
recover data sets from a DFSMShsm incremental backup version or from
a DFSMShsm dump
copy. DFSMShsm automatically
chooses the more recent copy of the data set if you allow users to
perform data set restores (see Controlling restoring of individual data sets from dump tapes).
There are special considerations for full volume restores of volumes
that were dumped using the BACKVOL DUMP command and whose data sets
contain RLS attributes.
DFSMShsm support activities
The following topics introduce the activities necessary to maintain
DFSMShsm-owned data sets and volumes. In addition to the activities
that cause the DFSMShsm program to perform its space management and availability management
functions, certain other activities maintain the security and efficiency
of DFSMShsm itself.
For example, backup copies of the control data sets should be kept,
and the DFSMShsm-owned volumes must be maintained.
Data recovery scenarios
Data recovery scenarios, describes real-life examples of data loss
and recovery situations. It describes a variety of situations where
data is lost, and what steps to take to recover that data. It also
includes situations where the lost data is a control data set or journal.