<br><br><div class="gmail_quote">On Mon, Feb 1, 2010 at 1:04 PM, Garrick <span dir="ltr"><<a href="mailto:garrick@usc.edu">garrick@usc.edu</a>></span> wrote:<br><blockquote class="gmail_quote" style="border-left: 1px solid rgb(204, 204, 204); margin: 0pt 0pt 0pt 0.8ex; padding-left: 1ex;">
On Mon, Feb 01, 2010 at 12:03:54PM -0500, Glen Beane alleged:<br>
<div><div></div><div class="h5">> On Mon, Feb 1, 2010 at 11:45 AM, Garrick <<a href="mailto:garrick@usc.edu">garrick@usc.edu</a>> wrote:<br>
><br>
> > The purpose of "complete" is so the scheduled can read why a job has<br>
> > exited. For that end, the current behavior is correct.<br>
> ><br>
><br>
> I can see a running job that gets qdeled end up in the "complete" state so<br>
> the scheduler can see why the job exited, but doesn't it seem strange to<br>
> qdel a job that is in the Q state and have it go to the C state?<br>
<br>
</div></div>It's for the scheduler guys, so it's up to them.<br>
<br></blockquote><div><br><br>fair enough <br></div></div>