Re: [BProc] mpirun

[Nicholas H. <he...@se...>]

On Tue, 22 Oct 2002 15:33:16 -0400
Nicholas Henke <he...@se...> wrote:

I seem to be having a ton of these errors on our testcluster, but not a
different cluster. I have tried to find the discrepencies, but no luck.
Do you have any idea what might be causing this ?

$ strace mpirun -d -p --np 2  ./cpi

[snip]
rt_sigprocmask(SIG_SETMASK, [], NULL, 8) = 0
nanosleep({5, 0}, {5, 0})               = 0
accept(3, {sin_family=AF_INET, sin_port=htons(33105),
sin_addr=inet_addr("192.168.2.10")}}, [16]) = 4
read(4, "\3048\0\0", 4)                 = 4
read(4, "P\201\0\0", 4)                 = 4
accept(3, {sin_family=AF_INET, sin_port=htons(32976),
sin_addr=inet_addr("192.168.2.11")}}, [16]) = 5
read(5, "\3058\0\0", 4)                 = 4
read(5, "\317\200\0\0", 4)              = 4
write(1, "  0    0 14532 192.168.2.10    3"..., 37  0    0 14532
192.168.2.10    33104
) = 37
write(1, "  1    1 14533 192.168.2.11    3"..., 37  1    1 14533
192.168.2.11    32975
) = 37
write(4, "\2\0\0\0", 4)                 = 4
write(4, "\0\0\0\0", 4)                 = 4
write(4, "\3048\0\0", 4)                = 4
write(4, "\300\250\2\n", 4)             = 4
write(4, "\201P", 2)                    = 2
write(4, "\3058\0\0", 4)                = 4
write(4, "\300\250\2\v", 4)             = 4
write(4, "\200\317", 2)                 = 2
close(4)                                = 0
write(5, "\2\0\0\0", 4)                 = 4
write(5, "\1\0\0\0", 4)                 = 4
write(5, "\3048\0\0", 4)                = 4
write(5, "\300\250\2\n", 4)             = 4
write(5, "\201P", 2)                    = 2
write(5, "\3058\0\0", 4)                = 4
write(5, "\300\250\2\v", 4)             = 4
write(5, "\200\317", 2)                 = 2
close(5)                                = 0
wait4(-1, Process 0 on node1.internal.org
[WIFSIGNALED(s) && WTERMSIG(s) == SIGINT], 0, NULL) = 14533
--- SIGCHLD (Child exited) ---
write(2, "rank 1 pid=14533 exited with sig"..., 38rank 1 pid=14533
exited with signal 2
) = 38
wait4(-1, xm_14532:  p4_error: net_recv read:  probable EOF on socket: 1
Connection failed for reason: : Connection refused
Connection failed for reason: : Connection refused
[WIFSIGNALED(s) && WTERMSIG(s) == SIGPIPE], 0, NULL) = 14532
--- SIGCHLD (Child exited) ---
write(2, "rank 0 pid=14532 exited with sig"..., 39rank 0 pid=14532
exited with signal 13
) = 39
wait4(-1, 0xbffff9f8, 0, NULL)          = -1 ECHILD (No child processes)
munmap(0x40017000, 4096)                = 0
_exit(0)