Weird I just got about 100 of these messages again this morning.
Something with the setup Tim M has that is causing the MTT client to
not send the mpi_install_id (maybe not sending any mpi_install data).
I won't have time to look at this in the near term, but as you all
are sitting down in Paris you might want to take a look at what he
has setup to see if there are any glaring errors.
-- Josh
On Oct 1, 2007, at 12:04 PM, Jeff Squyres wrote:
No, AI should also send the MPI_Install info. I successfully use AI
for HP MPI and Intel MPI coll bakeoff testing...
On Oct 1, 2007, at 5:59 PM, Josh Hursey wrote:
I have recently been receiving a ton of errors like the following. I
traced it as best I could to an 'Already Installed' configuration
that Tim M is running on BigRed for the collectives bakeoff. Is it
possible that the AlreadyInstalled is not even trying to send any
MPI_Install information?
-- Josh
Begin forwarded message:
From: jjhur...@open-mpi.org
Date: October 1, 2007 11:49:34 AM EDT
To: jjhur...@open-mpi.org
Subject: MTT server error (user: iu)
Reply-To: jjhur...@open-mpi.org
get_mpi_install_ids(0):
-------------------------------
SELECT mpi_install_id,
compute_cluster_id,
mpi_install_compiler_id,
mpi_get_id,
mpi_install_configure_id
FROM mpi_install
WHERE
mpi_install_id = '0'
-------------------------------
-------------------------------
Invalid mpi_install_id () given. Guessing that it should be 0 .
-------------------------------
_______________________________________________
mtt-devel mailing list
mtt-de...@open-mpi.org
http://www.open-mpi.org/mailman/listinfo.cgi/mtt-devel
--
Jeff Squyres
Cisco Systems
_______________________________________________
mtt-devel mailing list
mtt-de...@open-mpi.org
http://www.open-mpi.org/mailman/listinfo.cgi/mtt-devel
