[Beowulf] Re: [PVFS2-users] new mpich/pvfs2

Robert Latham robl at mcs.anl.gov
Mon Jun 13 10:03:49 PDT 2005


On Sun, Jun 12, 2005 at 04:51:23PM +0200, Michael Gauckler wrote:
> We are happy for any hint how to avoid this or any
> timings/confirmation from people who are willing to test our little
> program on their system.

It's great that you've provided the test case for us.  Unfortunatley I
had to greatly reduce the working set size.  I don't have enough RAM to
run with the paramters you provided  w/o hitting swap.   When I reduce
transactions, attributes, timesteps, and samples, I don't see a crash
with 4 processors and either 1 or 2  pvfs2 servers.  

Since you know this test program better than I do, can you come up
with a set of parameters that provides a sort of minimum failure test?
If the test only fails for these parameters, that's also good to know
(and I'll try to find some systems with more memory I can test on).  

==rob

-- 
Rob Latham
Mathematics and Computer Science Division    A215 0178 EA2D B059 8CDF
Argonne National Labs, IL USA                B29D F333 664A 4280 315B



More information about the Beowulf mailing list