[MPICH] mpdboot errors

[Luiz Mendes]

HI all,

Well, i have posted this stuff some days ago, in that ocasion i hadnt tried
somethings to stress all possibilities.

After several attempts i come back here again to ask for this following
error;

When i use mpdboot, only two machines are correctly recognized. For example:

I have 6 PCS. with names from "fisio1" to "fisio6".

With Mpich1, all these PCS have comunicate with each other without problems.
When i try with MPICH2 i have problems.

for example:

I defined a file with hosts

later i input the following command:

mpiboot -n 3 -f hostfile

With error:

sometimes
mpdboot_fisio1 (handle_mpd_output 374): failed to ping mpd on fisio3; recvd
output={}

and another time
mpdboot_fisio1 (handle_mpd_output 374): failed to ping mpd on fisio5; recvd
output={}

As you see, i think that this is not a problem of comunication between PCS,
because with 2 PCS, MPDboot works correctly.
And the same PC that reports communication errors on some ocasions in
another situations works correctly.

What it could be?


Thanks in advance
Luiz Mendes
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.mcs.anl.gov/pipermail/mpich-discuss/attachments/20070202/b782c86f/attachment.htm>