[torqueusers] Job bounces from status R to status Q

Garrick Staples garrick at usc.edu
Thu Dec 15 12:56:24 MST 2005


On Thu, Dec 15, 2005 at 01:44:49PM -0500, Michael Barnes alleged:
> I also have seen the bouncing from R to Q and it also bounces from node
> to node when I have a user submit a job requesting only one processor
> and staging in a large file (probably a timeout issue in the prologue,
> that I cannot find).

stagein happens before prologue, so that should be unrelated.

 
> A workaround when this happens is to have the user complain first, then
> tell them to rcp the file manually in their script.
> 
> This is with an older version of PBS not torque.  I believe it is the
> latest PBS release.  I'm not sure if anybody else has seen this issue
> and/or if it has been fixed in torque, but I would love to find a real
> solution to this problem.

I don't recall anyone mentioning this specific problem, but data staging
in TORQUE is very modified from the original OpenPBS.  Since you are
experiencing this problem with something that isn't TORQUE, I'm not sure
what to suggest here.

-- 
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/20051215/c31b98bd/attachment.bin


More information about the torqueusers mailing list