Case 2: Damaged journal, undamaged CDS
A damaged journal is often the result of a hardware-caused error on the journal volume. DFSMShsm issues error message ARC0026E when it encounters an input/output (I/O) error in the journal data set. Message ARC0026E states “JOURNALING DISABLED DUE TO I/O ERROR. MIGRATION, BACKUP, DUMP, TAPECOPY, TAPEREPL, RECYCLE, ARECOVER, AUDIT, AND EXPIREBV HELD.”
Once the problem occurs, follow this procedure:
- Verify the DFSMShsm CDSs are structurally sound. Using IDCAMS, run EXAMINE against each of the CDSs. The procedure runs very quickly, and does not require that you bring DFSMShsm down. The sample JCL to perform this procedure (which follows) comes from z/OS DFSMS Access Method Services Commands.
//EXAMCDS JOB //STEP1 EXEC PGM=IDCAMS //SYSPRINT DD SYSOUT=A //SYSIN DD * EXAMINE - NAME(?hlqual.MCDS) - ERRORLIMIT(0) //STEP2 EXEC PGM=IDCAMS //SYSPRINT DD SYSOUT=A //SYSIN DD * EXAMINE - NAME(?hlqual.BCDS) - ERRORLIMIT(0) //STEP3 EXEC PGM=IDCAMS //SYSPRINT DD SYSOUT=A //SYSIN DD * EXAMINE - NAME(?hlqual.OCDS) - ERRORLIMIT(0)
If there are no errors reported by EXAMINE, continue to the next step.
If errors are reported by EXAMINE, do not proceed to the next step. Go to Case 1: Damaged CDS, full journal.
- Stop DFSMShsm on all images.
- Do not copy the journal.
- Rename the current journal.
- Allocate a new journal with the old name. The following JCL allocates
a new journal with the old name.
//JRNLDEF JOB //STEP1 EXEC PGM=IEFBR14 //JRNLDD DD DISP=(,CATLG),UNIT=?3390,VOL=SER=?JNLVOL, // SPACE=(CYL,(100),,CONTIG),DSN=?hlqual.JRNL
- Start DFSMShsm in EMERG=YES mode, a proclib option.
- Issue the DFSMShsm BACKVOL CDS command to keep your CDSs in synchronization with the new journal.
- Delete the old journal.
- Issue the DFSMShsm SETSYS NOEMERGENCY command and then the DFSMShsm RELEASE command to resume normal processing.