The recovery process usually kicks in when Data Services jobs fail. A failed job, in most cases, means that some part of it has completed successfully and some part has not. The job which has failed right at the very beginning is rarely a problem and is of hardly any concern for recovery as all you have to do is to start it again.
Complications arise when the job fails in the middle of the insert, into a target table for example. Cases like that require you to either consider deleting already inserted records or even recovering a copy of the table from a backup using database recovery methods.
Recovery and error handling is an important part of robust ETL code. In this recipe, we will take a look at the methods used to develop ETL in Data Services and the functionality available in the software to make sure that the process of resuming failed processes goes as smoothly as possible.
The automatic job recovery feature available in Data Services does not...