DataSync should catch exceptions and mark the job as failed

This is a great feature request, thanks. I agree that DataSync should handle uncaught exceptions.

For the moment, what you can do is catch exceptions yourself and then return FALSE in your "phase function". DataSync will then call the job's fail_phase() method which of course you can override to do whatever you want.

dataync_scheduler module also has timeout functionality so if a job has been "running" for too long it will automatically call fail_phase() on it.

Please let me know if you think catching exceptions and running fail_phase() on the job would fulfill your request.

actually, let me know how this patch works for you.

oops, this one.

Pierre,

The reason the job restarts is because the default (and currently only) implementation of fail_phase restarts the current phase and queues it for re-run. It was always my intention to have 3 default implementations of fail_phase that you could choose from. Please test this patch with your application. It sounds like you will want to choose the "fail_job" method by overriding fail_phase in your job-type class and having something like:

function fail_phase($method = NULL) {
  return parent::fail_phase('fail_job');
}

Let me know how that works out for you.

Great. let me know how this works and hopefully we can release another alpha today, unless you have any other issues you've run into.

committed. thanks for testing