To reestablish access to previously deleted migrated data sets,
perform the following steps:
- Create a catalog entry for the data set. Use the IDCAMS DEFINE
NONVSAM command with the VOLSER(MIGRAT) and UNIT(tapeunit)
parameters.
- If the data set is NON-VSAM, run IDCAMS DEFINE data set to VOLSER
of MIGRAT to create the catalog entry.
Example JCL:
//IDCAMS JOB ,CLASS=A,MSGCLASS=H
//STEP EXEC PGM=IDCAMS
//SYSPRINT DD SYSOUT=*
//SYSIN DD *
DEFINE NONVSAM -
(NAME(dsname) -
DEVT(device type) -
VOL(MIGRAT))
/*
- If the data set is VSAM, use IDCAMS DEFINE base and data component
to VOLSER of MIGRAT to create the catalog entry.
Example JCL:
//IDCAMS JOB ,CLASS=A,MSGCLASS=H
//STEP EXEC PGM=IDCAMS
//SYSPRINT DD SYSOUT=*
//SYSIN DD *
DEFINE NONVSAM -
(NAME(xxxxx.ksds) -
DEVT(device type) -
VOL(MIGRAT))
DEFINE NONVSAM -
(NAME(xxxxx.ksds.data) -
DEVT(device type) -
VOL(MIGRAT))
/*
Note: For a VSAM data set, define the NONVSAM entry
for the base and data component.
- The MCD record must be altered, if it exists, or created, if it
does not exist.
- Use the following command to display the MCD record, if it exists:
FIXCDS D data.set.name DISPLAY
Then
issue the ASSIGNEDBIT command to turn on the bit, which indicates
that the data set has migrated:
FIXCDS D data.set.name ASSIGNEDBIT(ON)
- If the MCD record does not exist (the FIXCDS DISPLAY command failed),
the following AUDIT command will create it. The ML2 volser is needed
for the audit command. If you do not know the volser, check in one
of your journal backup versions to locate the journal entry of the
MCD record for the migrated data set. Note the migration volume serial
number, which is found at offset X'40' in the MCD record.
Run the following command with the volser parameter of the ML2 volume:
AUDIT MEDCTL VOLUMES(ML2volser) FIX SERIALIZATION(DYNAMIC)
- For VSAM data sets, if AUDIT issues ERR169 with the data component
name, then the MCD was created with the data component name instead
of the base cluster name. As explained in the ERR169 message, change
this name to the base cluster name before you perform a recall. Issue
the following command to change it:
FIXCDS D data.component.name NEWKEY(base.name)
- If the data set spans across tape volumes, you must repeat the
AUDIT MEDCTL command for each ML2 volser and issue the following FIXCDS
commands for the MCD record:
- Expand the MCD record:
FIXCDS D data.set.name EXPAND(z)
where z = number of volumes to which the data set spans ×
6
Example: Data set starts on volume A and spans to B and C; z = 2 × 6 = 12.
(2 = additional volumes, 6 = 6 characters for each volser)
- Add additional volsers to the MCDAVSN array at offset X'194'
FIXCDS D data.set.name PATCH(X'194' volserxvolsery)
where volserx =
the first additional volume and volsery equals the second
additional volume.
- Alter MCDNVSN with the number of additional tape volumes and MCDNVSNO
with the offset of the list of volsers:
FIXCDS D data.set.name PATCH(X'D4' X'zzzz0194')
where zzzz = the number of volsers and
0194 equals the offset to the list of volsers as determined in the
note below the following example.
Example: An
example command altering the MCDNVSN
FIXCDS D data.set.name PATCH(X'D4'X'00020194')
Note: To determine if a data set has spanned to additional volumes,
inspect the MCD record found in the journal backup. The number of "spanned
to" volumes is found in a two-byte field starting at offset X'D4'.
The location of the list of "spanned to" volumes is found at
offset X'D6'. Offsets are given from the beginning of the
MCD record. The MCD record found in the backup copy of the journal
is appended at the beginning by a 10 byte MLOG, a 44 byte MCK, and
a 20 byte MCH header record.
- Recall the data set.
- If you follow all steps correctly and you are still unable to
recall the data set, contact IBM® Support.