Re-Running Failed Jobs

Re-running a job is useful when a job was running correctly and was terminated for some (known or unknown) reason, such as a power failure or something else unexpected. It is not designed for jobs that fail because of configuration problems (i.e., component configuration errors).

Re-running a job uses the database to replay all the prior run data up to the point where the previous job terminated; then, the job continues to execute normally. Re-running a job allows the job to resume execution as quickly as possible by restoring the data from all the prior runs in the database. Isight does not attempt to re-execute failed runs if a process component finished with a non-failure completion code. Therefore, if the prior job ran for some time and then many components started failing, re-running the job will also lead to failed runs for only those components whose parent component finished successfully.

Once a process component fails, all its failed subcomponents are re-executed (i.e., no database lookup will be performed). If you re-run a job where a process component fails, Isight resumes execution where the failure began and continues running the job.