Rubén,

I'm quite serious when I use that phrase. Our computer systems are
subject to the forces of nature and to the physical reality in which
they live,
and there may be power glitches caused by storms, electrical events
that are disruptive, temporary failure due to corrosion of parts, disk
media failure, and even stray cosmic rays. These can be one-time,
non-repeatable effects that could cause a processing failure.

So maybe what I'm asking for is a "Terminator" service. Something
which will look and see how long a job has been in the system, how
much CPU it's using, the last time it changed state, etc., and makes a
decision to terminate it.

Obviously there could be an administrative tool to do this, but
administrators tend to be very busy doing other things and usually
aren't just sitting around caretaking a server.

Hank

2012/2/24 Rubén Pérez <[email protected]>:
> Totally agree, Hank (the acts of God part really made me laugh).
>
> The problem is that few people know their way around the workflow/conductor
> part of the software, but it would certainly be a nice feature if someone
> with enough knowledge can take the task.
>
> Just to share my opinion, I think that a mechanism to force-delete a
> workflow instance, rather than a 'self-destruct time' would be better. But I
> guess that's an open question.
>
> Best regards
_______________________________________________
Matterhorn-users mailing list
[email protected]
http://lists.opencastproject.org/mailman/listinfo/matterhorn-users

Reply via email to