Steps for preventing a full journal (damaged CDS)

To prevent a full journal, perform the following steps:

1. Ensure that no reorganization occurs if DFSMShsm is running on any image in the complex. You must do two things to accomplish this.
   • Make sure that SYSDSN enqueues are in the system's inclusion list for global resource serialization (GRS). This allows the SYSDSN enqueues to be passed to all images in the complex that share the CDSs.
   • Modify the JCL used for reorganization of the CDSs by using a disposition of OLD whenever a DD statement refers to a CDS.
2. Define the DFSMShsm CDSs with VSAM SHAREOPTIONS(2 3). You can find information about defining VSAM SHAREOPTIONS for DFSMShsm in the topic about VSAM SHAREOPTIONS for control data sets and associated CDS considerations in z/OS DFSMShsm Implementation and Customization Guide
3. Implement a procedure where the backup of the CDSs is closely monitored.

   One method of monitoring the backups of the CDSs is to review the CDS backup and export messages in the DFSMShsm backup activity log. Specifically, review the number of records processed by the export function in message IDC0005I, for each CDS. Compare the record count of the current CDS backup with the record count from the preceding days. The record count may not be the same, but it should be close.

   Normal update and deletion activity causes some fluctuation in the number of records in the CDSs. However, when a large decrease in records is detected, you should be able to account for the large difference.

   Some large fluctuations in the number of CDS records are normal, and result from processes that reduce the number of records in a CDS. For example, running the EXPIREBV command deletes records in the BCDS. DFSMShsm secondary space management deletes records in the MCDS. However, some large fluctuations in the number of records in a CDS can indicate that the CDS has been damaged. For example, in a multiple-MVS-image environment, the reorganization of a CDS occurring on one image while DFSMShsm is running on another image can damage the structure of the CDS, and cause a large fluctuation in CDS records.

   To eliminate any doubt about the soundness of the CDS structure, run the IDCAMS EXAMINE command against the CDS in question. Running the EXAMINE command against your CDSs every day minimizes the chance of CDS damage going undetected.