Restarting sequence jobs

Sequence jobs record checkpoint information so that you can restart some sequence jobs if they fail. If you run a restartable sequence and one of the jobs fails, you can fix the problem, then run the sequence job again from the point at which it failed.

About this task

You can enable or disable checkpoint recording at a project-wide level, or for individual sequence jobs. If you enable checkpoint recording for a sequence job, you can specify that checkpoints for individual components within a sequence job are not recorded. These components are forced to run whenever the sequence is restarted, regardless of whether they were run successfully already.

You can also specify that the sequence job handle automatically any errors that occur when the sequence job runs. By handling errors automatically, you do not have to specify an error handling trigger for every activity in the sequence job. You can enable or disable this setting on a project-wide basis, or for individual sequence jobs.


  1. Ensure that the job is able to be restarted. If a sequence job is restartable, and one of its jobs fail during a run, the following status appears in the Director client:
  2. Choose one of the following actions for your sequence job.
    Option Description
    Run job The sequence runs again using the checkpoint information to ensure that only the required components are run.
    Reset job All checkpoint information is cleared. When you run the sequence job again, the entire sequence job runs.
    Important: If the sequence job flow diverts to an error handling stage, InfoSphere® DataStage® does not record additional checkpoint information. This behavior ensures that activities in the error handling path are not be skipped if the sequence job is restarted and another error is encountered.