Backup and recovery of the IMS catalog

For certain IMS functions, such as the IMS management of ACBs or database versioning, the ability to recover the IMS catalog from back up copies is required. If these functions are not used, you can recover the IMS catalog by using one of the population utilities to rebuild it from ACB, DBD, and PSB libraries.

One of the functions that requires backup copies for recovery is the IMS management of ACBs. When IMS manages ACBs, the IMS catalog is the only repository for ACBs, DBDs, and PSBs in the IMS system and contains the only record of which of those resources are active.

Another function that requires backup copies for recovery is database versioning. When database versioning is used in IMS systems that use an ACB library, the DBDs for previous versions of a database exist only in the IMS catalog and cannot be repopulated from the ACB library. Only the version of the database definition (DBD) that is designated as active can be repopulated from the active ACB library.

IMS catalog data sets

The IMS catalog is a HALDB partitioned HIDAM (PHIDAM) database with a primary index data set, an indirect list data set (ILDS). The IMS catalog includes a HALDB partitioned secondary index (PSINDEX) database.

If the IMS management of ACBs is enabled, the IMS catalog also uses one or more IMS directory data sets. If any database or program view changes are pending activation, the IMS directory data sets might also include a staging data set.

Coordinating IMS catalog recovery with the active ACBs

In IMS systems that use IMS-managed ACBs, when you recover the IMS catalog from backup image copies, you must ensure that the DBD and PSB resource instances that are flagged as active in the recovered IMS catalog match the active ACBs in the IMS directory data set.

One way to ensure that the active instances in the recovered IMS catalog match the active ACBs in the IMS directory data set is to rebuild the IMS directory by using the IMS Catalog Directory Recovery utility (DFS3RU00).

However, because rebuilding IMS directory data sets can take a long time, consider creating backup copies of the IMS directory data sets that correspond to the image copies you create of the IMS catalog.

In IMS systems that use ACB libraries, when you recover the IMS catalog from backup image copies, you must ensure that the recovered IMS catalog contains the DBD and PSB instances that match the ACBs in the active ACB library. The time stamps of the records in the IMS catalog must match the time stamps of the corresponding ACB members in the active ACB library.

DBRC, the IMS recovery utilities, and the IMS catalog

All standard IMS utilities can run on the catalog data sets, including the image copy and database recovery utilities that are provided with IMS.

When the IMS catalog is managed by DBRC, the utilities that are provided with IMS create recovery information in the log data sets when the IMS catalog is updated. DBRC manages the logs, image copies, and JCL required for recovery of the IMS catalog. You can perform a full database recovery or a point-in-time recovery for the IMS catalog partitions.

When DBRC is not used with the IMS catalog, the IMS catalog can be recovered by using the standard backup and recovery processes that are used for other HALDB databases. However, you must have processes in place to manage the logs, image copies, JCL, and so on.

Recovery by repopulation

If your installation does not use either the IMS management of ACBs or database versioning, as an alternative to recovering the IMS catalog from image copies and log records, you can re-create the IMS catalog from your ACB, DBD, and PSB libraries by running either the IMS Catalog Populate utility (DFS3PU00) or the ACB Generation and Catalog Populate utility (DFS3UACB) to reload the IMS catalog.

Re-creating the IMS catalog from ACB, DBD, and PSB libraries by using one of the population utilities restores only the record segments in the IMS catalog that are in the libraries. Use this method only if you do not need record segments for past instances of your DBDs and PSBS or the historical metadata that those segment instances contain.

An initial load of the catalog by either of the population utilities does not create any database recovery information in the log data sets. Create an image copy immediately after the IMS catalog is loaded to ensure that the image copy is consistent with the active ACB library.

Recovering the IMS catalog secondary index

In a data-sharing environment, recovering the secondary index is required in a full recovery for the IMS catalog. Recover the secondary index by using an image copy. Besides, you can also rebuild the secondary index by using IMS Tools, such as IMS Index Builder.

Recovering the IMS catalog directory data sets

When the IMS management of ACBs is enabled, you can use the IMS Catalog Directory Recovery utility (DFS3RU00) to recover the IMS directory data sets from database image copies after the recovery of IMS catalog. The IMS Catalog Directory Recovery utility rebuilds the IMS directory data sets and recovers the ACBs of IMS databases and programs to the directory by reading active resources from the IMS catalog.