Restoring an aggregate with DFSMSdss logical restore

Use DFSMSdss logical restore to restore a zFS aggregate. If the original aggregate (in the example, hlq.ZFS.AGGR004) still exists, the aggregate is restored into a new aggregate (in the example, OMVS.PRV.AGGR005.LDS0005). Figure 1 is an example of a job to restore a zFS aggregate.

Figure 1. Job to restore a zFS aggregate

//ZFSREST1 JOB (OS390),'PROGRAMMER',CLASS=A,
// MSGCLASS=X,MSGLEVEL=(1,1)
//*-----------------------------------------------------------------
//* THIS JOB RESTORES A ZFS AGGREGATE.
//*-----------------------------------------------------------------
//ZFSREST EXEC PGM=ADRDSSU,REGION=0M
//SYSPRINT DD SYSOUT=*
//SYSABEND DD SYSOUT=*
//INDS DD DISP=SHR,DSN=hlq.AGGR004.BACKUP
//SYSIN DD *
 RESTORE DATASET(INCLUDE(**)) -
 CATALOG -
 RENAMEU( -
(hlq.ZFS.AGGR004, -
OMVS.PRV.AGGR005.LDS0005) -
) -
WRITECHECK -
INDD(INDS)

Leading blanks are required before the control statements (RESTORE, CATALOG, RENAMU).

For a compatibility mode aggregate, perform the following steps after the aggregate is restored:

Unmount the original aggregate (in this case, hlq.ZFS.AGGR004) if it still exists (this also detaches it).
Mount the file system in the restored aggregate (in this case, OMVS.PRV.AGGR005.LDS0005).

Figure 2 is an example of a job to perform a logical restore of a zFS aggregate using DFSMSdss by replacing the existing aggregate. The backup is restored into the original aggregate (in this case, hlq.ZFS.AGGR004). The aggregate cannot be mounted (or attached) during the restore operation.

Figure 2. Job to restore a zFS aggregate with replace

//ZFSREST2 JOB (OS390),'PROGRAMMER',CLASS=A,
// MSGCLASS=X,MSGLEVEL=(1,1)
//*---------------------------------------------------
//* THIS JOB RESTORES A ZFS AGGREGATE.
//*---------------------------------------------------
//ZFSREST EXEC PGM=ADRDSSU,REGION=0M
//SYSPRINT DD SYSOUT=*
//SYSABEND DD SYSOUT=*
//INDS DD DISP=SHR,DSN=hlq.AGGR004.BACKUP
//SYSIN DD *
 RESTORE DATASET(INCLUDE(hlq.ZFS.AGGR004)) -
 CATALOG -
 REPLACE -
 WRITECHECK -
 INDD(INDS)

Leading blanks are required before the control statements (RESTORE, CATALOG, RENAMU).

For more information about DFSMSdss logical restore, see z/OS DFSMSdss Storage Administration.

Beginning in z/OS® V1R2, zFS enhanced its support for the backup change activity flag in the VTOC (D1DSCHA in the Format 1/8). This flag indicates to a program (like DFSMShsm) whether the backup of a file system is needed (that is, data in the file system has been modified since the last backup).

In releases prior to z/OS V2R1, zFS would set the change activity flag when a file system was mounted. This is no longer done. Essentially, zFS will cause the setting of the change activity bit in the following cases:

During the first write after a MOUNT
During the first write after a successful backup (that is, after a successful reset of the change activity flag)
During log recovery (that is, during the replay of an aggregate log during the next mount after a system failure)
During salvager operation if the log is replayed or a repair is made

The formatting of a new zFS aggregate will not cause the setting of the change activity flag. If an existing zFS aggregate is formatted using the -overwrite option, then the change activity flag will be set.

z/OS V2R1 zFS will also supply an application programming interface, which can be used to reset the change activity flag for a file system. This interface is intended to be used by DFSMSdss when doing a backup of a mounted zFS file system. For more information, see Reset Backup Flag.