[torqueusers] Removing the "exec_host" attribute from a queued job ?

Garrick Staples garrick at usc.edu
Tue Sep 20 12:46:33 MDT 2005


On Tue, Sep 20, 2005 at 08:41:55AM -0400, Andrew J Caird alleged:
> I'm not sure if this is exactly the same thing, but we see this in maui 
> sometimes; it looks like:
>    Allocated Nodes:
>    [nyx020:1]
> but if nyx020 has crashed, it just sits there.  We use the maui command:
>    runjob -c <jobid>
> According to the help for runjob:
>    [ -c ] // CLEAR (clear stale job attributes)
> and this seems to prompt maui to relook at where the job should be run.

That we need that is dumb.  The scheduler should really handle this
better.

If someone can figure out the details of when this problem happens, please
submit a bug report.

-- 
Garrick Staples, Linux/HPCC Administrator
University of Southern California
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://www.supercluster.org/pipermail/torqueusers/attachments/20050920/bc9607fd/attachment.bin


More information about the torqueusers mailing list