8.4

examples would be W3813, W3830, W3991, W4005. These correspond to issues such as not
having any space to write files, permission errors on folders or files, etc.
Creating and Using Error Processes
An Error Process is a special type of process that never runs on its own, and cannot be called
using the GoSub or Send to Process tasks. It can only be used in the On Error tab of a task in
your process, and will be triggered if the Send to Process option is checked in that tab and an
error process is selected in the drop-down list.
To create an error process, simply replace the initial input task by the InputErrorBin input task,
and that process automatically becomes able to handle error jobs sent to it. It is up to you,
however, to decide how that error job will be handled.
For example, you could place the job file in a specific folder, then send an email to a supervisor
indicating that a job has failed. Or you could update a database with an error status so that it
appears on a customer's online order. You could also zip the order up and send it to an
administrator, while simultaneously advising the person that sent the job that it failed.
You can have as many error processes as you can normal processes - that is, you are limited to
512 processes, subprocesses, startup processes and error processes combined.
The following information is available from within your information process when it is
triggered:
l Job Information variables (%1 to %9)
l The data file as it was before starting the task
l Global variables (which are, of course, available anywhere)
l Aseries of variables containing information about the error, the task that triggered it and
the process that contained it. See "Standard Variables" on page 470
Note
Local variables in the process are not sent to error processes, even if the error process has a variable
of the same name.
Page 46