The effect and recovery method of abnormal abort during catalog backup

During the catalog backup recovery process, the import job terminates abnormally under certain conditions. At this point, catalog will be in a damaged state: The original catalog can not continue to use, import job can not continue. This article mainly introduces an example of such a scene, then restores and proposes how to avoid this anomaly.

Catalog for the Zos system is self-evident, so we often need to back up catalog or in the reorganization (reorganize) catalog (because the VSAM file has 123 partition limits, so when the total number of partitions to reach a certain proportion, will need to reorganize catalog). It is important to ensure that sufficient memory is allocated during the export/import of the catalog, otherwise it is likely that the job will terminate abnormally (usually ABEND878, Getmain ERROR). This paper describes in detail such a scenario and the destructive effect of the abnormal termination of the operation on catalog, and restores the catalog successfully.

The example described in this article is presented in the process of reorganize the catalog. There are no errors in the process of checking and lock the catalog. The same export process is also very smooth, but in the catalog import (import), the job abnormally terminated, after inspection found that the operation is not enough memory (abend878), in the EXEC statement has region=4m, So there are abend878 (see output below).

Ief403i catipt-started-time=03.44.58

iea705i ERROR DURING getmain SYS CODE = 878-10 catipt IMPORT 00

iea995i Symptom DUMP OUTPUT 382

iea705i 00f94d00 007ff3a0 007ff3a0 00001200 00000ae0

SYSTEM Completion code=0c2 REASON code=00000002

time=03.45.00 seq=37883 cpu=0086 asid=003e

The following issues should give you special attention. In general, increasing memory will solve the problem of job running. So after you specify region=0m, and then run the job again, but you still get a catalog error, you cannot import the backup catalog. Then try unlocking and using the original catalog to get the following error message (Note: Return code 246 only indicates that catalog has an internal error and cannot help you solve more problems).

ALTER CATALOG. PLEX2. MDM UNLOCK 00070

Idc3014i CATALOG ERROR

idc3009i * * VSAM CATALOG return code was 246-reason code is igg0clha-0

Idc0532i **entry CATALOG. PLEX2. MDM not altered

Idc0001i FUNCTION COMPLETED, highest CONDITION CODE was 8

Here is the process of trying to fix the catalog. Before the repair was started, it was not realized that the catalog had been damaged, so it was repaired in accordance with the general process. In fact, it is not possible to use the original catalog or to restore catalog in a normal way. Catalog has been placed in the deleted state (below) by looking at the catalog address space. Any operation of the catalog at this time is prohibited.

-F catalog,allocated

iec351i CATALOG Address spaces MODIFY COMMAND ACTIVE

Iec348i Allocated Catalogs

*cas*******************************************************************

* Flags-volser-user-catalog NAME% *

* ysi-r-p2wt15 0001 CATALOG. PLEX2. WIT 1 *

* ysd-r-p2ss32 0001 CATALOG. PLEX2. MDM N/A *

Then I try to refresh the catalog in a disconnect/connect way, trying to reset the catalog state in this way. Then, after importing (import) catalog, when you define alias for the new catalog, you find that catalog is still inaccessible. After displaying the address space, it was found that it was still placed in the deleted state. Then, I choose to restart the catalog address space (this is a great impact on the system of the action, must choose a suitable time window), trying to refresh the state of catalog again, still in vain.