[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Condor-users] File transfer timeout

Hello, all.

I'm attempting to submit a test MPI job with about 4 GB of data.  I'm
using Condor-6.7.17 with mpich-1.2.4.  I'm able to submit the job,
which is scheduled and dispatched.  I can then cd to the
~condor/execute/dir_pid directory and watch the file transfer.  The
nodes are dual processor, so I get two vms on each node.  Inevitably,
the first vm gets its files and the second times out waiting for the
transfer to begin.  It appears that the transfers are sequential,
rather than in parallel to all the nodes.

Does anyone have a solution or advice on this problem?  I've tried
compressing the input data, which reduces it to about 1.2 GB and
adding a step to mp1script to extract it.  Same result.