z/OS DFSMS OAM Planning, Installation, and Storage Administration Guide for Object Support
Previous topic | Next topic | Contents | Contact z/OS | Library | PDF


CYCLE START TIME and CYCLE END TIME

z/OS DFSMS OAM Planning, Installation, and Storage Administration Guide for Object Support
SC23-6866-00

As described in Understanding storage management cycles, the storage management cycle ensures that every object scheduled for processing is placed in the correct level of the object storage hierarchy, is deleted, expired, or backed up, and, if necessary, is flagged for action during a later storage management cycle. There are five methods by which management cycles can be controlled:
  • Manual Start (All Groups)

    You can start the storage management cycle manually for all storage groups by using the MODIFY OAM, START, OSMC operator command. START OSMC starts all groups that either have the system name or no system name specified as the OSMC processing system. (See Starting OSMC functions for command syntax.)

  • Manual Start (Individual Group)

    You can start the storage management cycle manually for an individual storage group by using the MODIFY OAM,START,STORGRP operator command. Using this approach, you can directly control the processing sequence, such as using this technique to give priority to a storage group with many objects.

  • Manual Stop

    You can stop the storage management cycle manually by using the MODIFY OAM,STOP,OSMC command. (See Stopping OSMC for command syntax.)

  • Manual Stop (Individual Group)

    You can stop the storage management cycle manually for an individual Object or Object Backup storage group by using the MODIFY OAM,STOP,STORGRP operator command.

  • Cycle Start Window

    You can start the storage management cycle automatically for Object and Object Backup storage groups by specifying CYCLE START TIME and CYCLE END TIME parameters in the storage group definitions. If using the automatic startup in an OAMplex, you should specify an OSMC processing system name to avoid multiple starts for the same storage group on different systems. If you start OSMC during the window delimited by those times, the storage management cycle is started for that storage group. This is the usual method for controlling storage management cycles. You can also indicate that no automatic processing for the storage group is to be performed by specifying NONE for the CYCLE START TIME and leaving the CYCLE END TIME blank.

    The SETOSMC CYCLEWINDOW keyword interacts with the CYCLE START TIME and CYCLE END TIME parameters in the following ways:
    • If SETOSMC CYCLEWINDOW is not specified, STARTONLY is the default mode for the cycle start window.
    • If you specify the SETOSMC CYCLEWINDOW(STARTONLY) keyword in the CBROAMxx member, CYCLE START TIME and CYCLE END TIME describe a window during which the storage management cycle might start. These cycle times do not define the length of the processing period. The storage management cycle might continue to run after the specified CYCLE END TIME has passed. Consider adjusting the start time or times if the processing for one group extends into the start time for another; otherwise, resource contention can become severe enough to affect the total time that it takes to perform storage management cycle processing for all groups.
    • If you specify the SETOSMC CYCLEWINDOW(STARTSTOP) keyword in the CBROAMxx member, CYCLE START TIME and CYCLE END TIME describe a window during which the storage management cycle starts and ends. The storage management cycle ends at the specified stop time. Although OSMC stops processing new work for this storage group, this group can continue to finish what work it has started.

    Related reading: For more information on the SETOSMC CYCLEWINDOW keyword, see SETOSMC keyword definitions.

All Object or Object Backup storage group definitions must define a window where the storage management cycle starts for the storage group, or indicate that no automatic processing be performed for the storage group. Consider the following issues as you select window start and end times for each storage group:
  • Number of optical drives
  • Number of tape drives
  • Number and size of objects moving through the hierarchy
  • Backup requirements
  • Time required to process the group
  • Impact on end users who might be doing retrievals
  • Storage management cycles for other groups
  • Application usage patterns
  • General maintenance operation requirements
Processing during the storage management cycle for a group does not require use of an optical or tape drive under the following conditions:
  • The storage group does not specify a library.
  • There is no class transition that requires moving an object to optical or tape storage.
  • No objects require backup.

Storage management cycle processing requires at least one drive if any objects are moved or backed up to optical or tape storage. (See DRIVE STARTUP THRESHOLD and TAPEDRIVESTARTUP on page TAPEDRIVESTARTUP(threshold in megabytes) for other considerations.) If a storage management cycle is in process on more than one Object or Object Backup storage group at a time and the number of groups exceeds drive availability, frequent volume mounts occur in an attempt to satisfy the requests to write objects to optical and or tape volumes. For example, when objects are written to a mounted volume for one group, that volume must be demounted to allow the mounting of another volume to accept the objects for a different group. Unless you limit resource consumption during the storage management cycle by some other means, you must not specify overlapping start windows for more groups than you have drives. (See OAM cataloged procedure parameter (MAXS).)

If an object requires more than one backup copy, the first and second backup copies are written to separate Object Backup storage groups. Objects are written to backup volumes in the Object Backup storage group specified in the SETOSMC statement for the Object storage group to which they belong, or, if a SETOSMC statement is not specified, the backup copies are written to backup volumes in the default Object Backup storage group specified.

More than one Object storage group can use the same Object Backup storage group for the same backup copy (first or second); therefore, first backup copies of objects from one group can reside on the same volume as first backup copies from other groups and second backup copies of objects from one group can reside on the same volume as second backup copies from other groups. These objects are written during the storage management cycle for the group containing the object. If some groups require object backup, but some do not, consider processing groups that require object backup concurrently with groups that do not require object backup.

Recommendation: Process the storage management cycle for an Object storage group while other activity for the objects in the Object storage group is light. For example, specify a cycle start window during a period when applications are not accessing data heavily. You must consider the effect of concurrent object use in a group during the storage management cycle for that group. DB2 performs deadlock detection on tables (directory tables in particular) that are shared by tasks performing the storage management cycle processing and by user tasks requesting OAM functions through the application interface. The potential for DB2 deadlocks is much greater if an application is accessing data in a group during the storage management cycle for that group.

Go to the previous page Go to the next page




Copyright IBM Corporation 1990, 2014