This defect has been raised because Testimony detected a different output in the batch job log.
If we look at the investigate screen, we can see that there were 7 rows in the job log in production, whereas there were only 4 rows in the log during the playback.
The first two rows in a job log are always the Job Started and Step Started entries, so these can be ignored in our analysis (and indeed, these are identical in both the recording and the playback).
Looking at rows 3 and 4, however, we can see that in the recording the batch job output some messages relating to purchase orders that it was processing, whereas in the playback the job output a “No suitable purchase requisitions found” message and then finished.
The “Job finished” message in the recording (which is always the last entry in the log for a job that ran successfully) can actually be found at row 7:
This is a classic example of a data-related failure. In this case, either because of script sequencing or the failure of one or more previous transactions, when this job ran during the playback there were no records that met the selection criteria for the job.