errno 32 errors

[Joshua Thomas]

Hello all. This is my first posting here.

I'm using MPICH 1.2.4, with the ch_p4mpd device, on a cluster of SMP Athlons.
Compilation and install went w/o any problems, and all the test programs ran 
without trouble. The first programs I wrote myself even ran without any 
problems.

Now, however, I've written something a bit more complex, and about 50% of the 
time it dies with:

p0_29708: (0.055913) net_recv failed for fd = 17
p0_29708:  p4_error: net_recv read, errno = : 104

or

p0_29713: (0.022796) net_send: could not write to fd=17, errno = 32
p0_29713:  p4_error: net_send write: -1

I did some google searching, and noted that at one point this might have been 
a SMP issue, and a patch had been provided, though I could not locate it.

Any help would be GREATLY appreciated!

-- 
Joshua F. Thomas
Research Assistant | Fiber Optics Manufacturing
University of Connecticut
Lab: 860-486-0624
thomasj at engr.uconn.edu
http://www.engr.uconn.edu/ofmrl/