[torquedev] Cpuset behavior in TORQUE 3.0

Garrick Staples garrick at usc.edu
Fri Jul 2 12:51:11 MDT 2010


On Fri, Jul 02, 2010 at 11:16:47AM -0600, David Beer alleged:
> In this discussion, it would be great if the original developer of TORQUE cpusets (I don't actually know who that is) could weigh in on the matter, and give us any special reasons why changing the behavior would be harmful.

*raises hand*

trunk]$ svn blame src/resmom/linux/cpuset.c  | head
  1101    garrick #include <pbs_config.h>
  1101    garrick 
   944    garrick #include <sys/stat.h>
   944    garrick #include <dirent.h>
  1613    garrick #include <stdio.h>
  1735    garrick #include <sys/types.h>
  1735    garrick #include <signal.h>
  1735    garrick #include <unistd.h>
  2146       glen #include <sys/param.h>
   944    garrick 
svn: Can't write to stream: Broken pipe

I wrote the original cpuset code as a starting point without actually having a
cluster with cpuset supporting kernels. I make no claims of suitability or
correctness.

My thought was to just use cpusets if they existed and not make a ruckus if
they didn't.

My real hope was to give people something on which to play. At the time noone
had all of the answers and I was making it up as a I went along. Hopefully now
someone has a better understanding of what needs to be done.

-- 
Garrick Staples, GNU/Linux HPCC SysAdmin
University of Southern California

Life is Good!
-------------- 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/torquedev/attachments/20100702/472571b8/attachment.bin 


More information about the torquedev mailing list