Controlling error tolerance in UNLOAD

The ERROR_THRESHOLD keyword can be used to enable the unload function to encounter and bypass a specified number of segments containing pointer errors without terminating.

When ERROR_THRESHOLD is used with the UNLOAD command, the database record is unloaded to the output file up to the point of the segment in error, and processing continues with the next accessible segment. The segment containing the pointer error (and its dependents) is bypassed, and is not written to the unload file. When the number of errors encountered in all database records that have been read exceeds the value specified on the ERROR_THRESHOLD keyword, processing of the area terminates. If an unload function terminates because of pointer errors, requested image copies will still be created.

When the default value of 0 (zero) is in effect for the ERROR_THRESHOLD keyword, the unload function terminates when the first pointer error is encountered.

The example in the following figure requests a DEDB unload and will bypass up to three segments containing pointer errors before terminating the unload function. The three segments containing pointer errors (and their dependent segments) are not written out, and unload processing continues with the next accessible segment. If a fourth pointer error is encountered the unload function terminates.

When you specify SDEP_PROCESS=PHYSICAL on the UNLOAD command, the process does not read SDEPs by following pointers. If you want the ERROR_THRESHOLD keyword to detect and bypass any pointer errors detected in the SDEP portion of the DEDB, then you must specify SDEP_PROCESS=LOGICAL.