After each DASD data set is written to tape successfully, a CA Disk index entry is placed in the DSNINDEX subfile and the delete and catalog actions are taken as needed. With this design, any tape created by CA Disk is a good tape, even though a system failure or job cancel can have occurred.
If your installation uses a tape library management system such as CA 1 or CA Dynam/TLMS, you must ensure that they do not designate such tapes as scratch tapes. EDM support was designed for this contingency. For more information, see the user exit description for TMSCLEX-Tape Management System Control Interface Exit in the Systems Guide or contact your tape management vendor.
If a job fails, the exact same archive or incremental backup command request can be resubmitted. CA Disk will request a new scratch tape and processing will continue where it left off. In reality, all volumes and all data sets are actually rescanned for selection, but if they were successfully selected and processed in the immediately preceding partial run, most selection types will not reselect them. For example, an unused data set that was archived and deleted will not be seen by the restarted job. Changed data sets backed up successfully will no longer have the change bit on and will therefore not be reselected, unless the NORESET parameter is also specified.
If an archive or backup fails because the data set is in use, the data set can be automatically queued for deferred processing by specifying sysparm ARCFDEFR with a value of Y. For more information on this sysparm, see ARCFDERn in the Systems Guide.
|
Copyright © 2015 CA Technologies.
All rights reserved.
|
|