Recovering a data set from a fast replication backup

You can use the FRRECOV command to recover individual data sets from a copy pool backup. On FRRECOV, specify one or more data set names (fully or partially qualified). If you specify a partially qualified data set name with a filter (wild card), DFSMShsm processes every cataloged data set in the standard search order that matches your filtering criteria. The FRRECOV command allows you to:

Table 1 shows the wildcard support on the FRRECOV command.

Table 1. Wildcard support on the FRRECOV command.
Type of data sets that are specified with the FRRECOV command Moved data sets Deleted data sets Original allocation
Single fully qualified data set Process Process Process
Multiple fully qualified data sets Process Process Process
Wildcard specified on a single or multiple data sets Process Not processed Process
Note: You cannot use wild cards for deleted data sets.

For a FRRECOV of a data set to be successful, DFSMShsm must have catalog information from either the catalog information data set captured at backup time or current catalog information at the time of recovery. If the catalog information is not captured, but the data set is cataloged at the time of recovery, the data set recovery is attempted. If the catalog information is not captured and the data set is cataloged at backup time, but it is not cataloged at the time of recovery, the individual data sets cannot be recovered using the DSNAME() keyword.

If you omit a copy pool catalog name from the copy pool definition, data sets from that catalog are eligible for recovery if they were backed up in the copy pool and if they are currently cataloged to the same volumes in which they were allocated to at the time of backup. If you specify a catalog name that does not belong to the copy pool, DFSMShsm captures the information for that catalog at the time of backup. During FRRECOV processing, DFSMShsm uses the irrelevant catalog information to attempt to recover the data set, determines that the data set is not backed up as part of the copy pool, and fails the request. You can identify data sets that are backed up as part of a copy pool version using the DATASETS keyword on the LIST COPYPOOL command.

When you attempt to recover a data set that has changed volumes, the data set might be scratched and re-cataloged or just re-cataloged. If the data set is cataloged at the backup time, but not cataloged at the recovery time, the data set is allocated to the same volume on which the data set was allocated at the time of backup and DFSMShsm re-catalogs the data set. If a data set is cataloged to different volumes at the recovery time than it was cataloged at the backup time, the data set is scratched, allocated to the volumes on which the data set was allocated at the time of backup, and then DFSMShsm re-catalogs the data set. If this initial recovery attempt fails, DFSMShsm will perform SMS volume selection within the primary storage group and retry data set recovery.

If you define a storage group in multiple copy pools, and then create a backup, the data sets in this storage group have multiple copies on the backup volumes. To recover these data sets, use the FROMCOPYPOOL(cpname) keyword on FRRECOV command to identify which copy of the data set to recover. You must also specify FROMCOPYPOOL(cpname) and a fully qualified data set name to recover a deleted data set. If you do not specify FROMCOPYPOOL for a deleted data set, DFSMShsm fails.

FRRECOV processes each data set independently. If a data set is not found, FRRECOV continues processing until the other specified data sets are processed. On completion, FRRECOV processing issues messages to indicate which data sets are recovered and which requests fail. For wait-type requests, FRRECOV returns control when all of the data set recoveries have completed.

For a specified data set name, the high-level qualifier can be fully or partially qualified. You cannot specify two asterisks ('**') as the high-level qualifier or as the only qualifier but it can be specified in place of subsequent qualifiers. You can specify '*' within the high-level qualifier as long as it is not the first character specified. The remainder of the data set name can include any arrangement of wildcard characters. These restrictions help to prevent inadvertent requests for large numbers of data sets.

You cannot recover the following types of data sets using fast replication data set recovery:
In the following example, DFSMShsm requests a fast replication recovery for two data sets (DSNAME1 and DSNAME2). 
FRRECOV DSNAME(DSNAME1,DSNAME2) REPLACE
You can use the following command to recover individual data sets: 
FRRECOV DSNAME(USER.DATA.**,USER.LOG.**,USER.JOURNAL) REPLACE

The DSNAME keyword is mutually exclusive with the TOVOLUME and COPYPOOL keywords. If GEN, VER, DATE, TOKEN, or FROMDUMP(DVOL) are not specified, then the default is GEN(0).

NEWNAME is an optional keyword that allows you to recover a data set to different name on the same or different system, while preserving the originally-named data set for analysis or other purposes. You must specify the DSNAME keyword if you are also specifying NEWNAME(newdsname). If a data set already exists with the same name as the new name you are specifying, the REPLACE keyword must be specified to replace the existing data set.

REPLACE is a required keyword for cataloged data sets. FROMCOPYPOOL is a required keyword for deleted data sets. In order to recover a data set is that is uncataloged but exists on a volume, you must specify both REPLACE and FROMCOPYPOOL.

FROMDASD and FROMDUMP are mutually exclusive optional parameters that can be used to indicate the source of the recovery. The default is to recover the data set from DASD. If a DASD copy is not available, then the tape copy is recovered. You cannot specify DVOL(dvol) with GEN, VER, DATE, or TOKEN.

FROMCOPYPOOL is an optional keyword, but you must specify this keyword if the volume or volumes that a data set resides on belong to multiple copy pools. FROMCOPYPOOL is only valid with the TOVOLUME and DSNAME keywords. You must specify FROMCOPYPOOL for recovery of deleted data sets. NOCOPYPOOLBACKUP is an optional keyword. When a partially qualified data set name is specified and a copy pool backup version is not found for a data set that matches the filtering criteria, the default return code is 8. When RC4 is specified with this keyword, a return code of 4 is returned instead. This keyword is ignored for fully qualified data set names.

FASTREPLICATION is an optional keyword. It overrides any value specified on the SETSYS FASTREPLICATION command. If not specified, the value of SETSYS FASTREPLICATION is used. This keyword is only valid with DSNAME. The command fails if this keyword is specified with COPYPOOL or TOVOLUME.

You can use the LIST COPYPOOL command to list all data sets that are backed up with a specified version. The list output also contains the data sets that were backed up but have been deleted, and indicates whether a full, partial, or no catalog information is captured for a specific copy pool version. Data sets are listed in alphabetical order by catalog: 
LIST COPYPOOL(cpname) DATASETS(GENERATION(gen)|TOKEN(token))

Related reading

Parent topic: Recovering a fast replication backup