GNU bug report logs - #74940
[PATCH] gnu: openmpi-4: Fix hardcoded call to ssh command.

Previous Next

Package: guix-patches;

Reported by: Romain GARBAGE <romain.garbage <at> inria.fr>

Date: Wed, 18 Dec 2024 09:45:02 UTC

Severity: normal

Tags: patch

Done: Ludovic Courtès <ludo <at> gnu.org>

Bug is archived. No further changes may be made.

Full log


Message #5 received at submit <at> debbugs.gnu.org (full text, mbox):

From: Romain GARBAGE <romain.garbage <at> inria.fr>
To: guix-patches <at> gnu.org
Cc: ludovic.courtes <at> inria.fr, Romain GARBAGE <romain.garbage <at> inria.fr>
Subject: [PATCH] gnu: openmpi-4: Fix hardcoded call to ssh command.
Date: Wed, 18 Dec 2024 10:37:39 +0100
This patch fixes a runtime error when launching an MPI program using
mpiexec/mpirun in an environment where the ssh command is not
available.

* gnu/packages/mpi.scm (openmpi-4)[inputs]: Add openssh-sans-x.
(openmpi-4)[arguments]: Fix path in code.

Change-Id: I65255186c37b6eaba3c283c7046163abc2cb38b7
---
The runtime error can be reproduced with:
---8<------------------>8---
$ guix shell -C openmpi <at> 4 hello -- mpirun -np 2 hello
--------------------------------------------------------------------------
The value of the MCA parameter "plm_rsh_agent" was set to a path
that could not be found:

  plm_rsh_agent: ssh : rsh

Please either unset the parameter, or check that the path is correct
--------------------------------------------------------------------------
[guix-A102:00001] [[INVALID],INVALID] FORCE-TERMINATE AT Not found:-13 - error plm_rsh_component.c(335)
---8<------------------>8---

The size of the closure as given by guix size is 406.7 MiB after the
patch vs 398.3 MiB before the patch.
---
 gnu/packages/mpi.scm | 11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)

diff --git a/gnu/packages/mpi.scm b/gnu/packages/mpi.scm
index 661531abb9..60a1952809 100644
--- a/gnu/packages/mpi.scm
+++ b/gnu/packages/mpi.scm
@@ -53,6 +53,7 @@ (define-module (gnu packages mpi)
   #:use-module (gnu packages parallel)
   #:use-module (gnu packages pkg-config)
   #:use-module (gnu packages python)
+  #:use-module (gnu packages ssh)
   #:use-module (gnu packages valgrind)
   #:use-module (srfi srfi-1)
   #:use-module (ice-9 match))
@@ -220,7 +221,8 @@ (define-public openmpi-4
                      gfortran
                      libfabric
                      libevent
-                     opensm)
+                     opensm
+                     openssh-sans-x)
                (if-supported psm)
                (if-supported psm2)
                (if-supported ucx)
@@ -291,6 +293,13 @@ (define-public openmpi-4
                            '("./ompi/mca/io/romio321/src/io_romio321_component.c")
                          (("MCA_io_romio321_COMPLETE_CONFIGURE_FLAGS")
                           "\"[elided to reduce closure]\""))))
+                   (add-before 'build 'ssh-absolute-path
+                     (lambda* (#:key inputs #:allow-other-keys)
+                       (substitute* "./orte/mca/plm/rsh/plm_rsh_component.c"
+                         (("mca_plm_rsh_component.agent = \"ssh : rsh\";")
+                          (format #f "mca_plm_rsh_component.agent = \"~a : rsh\";"
+                                  (search-input-file inputs
+                                                     "bin/ssh"))))))
                    (add-before 'build 'scrub-timestamps ;reproducibility
                      (lambda _
                        (substitute* '("ompi/tools/ompi_info/param.c"

base-commit: cd018214a3c0319451ab46855ea395b2be6a15dc
-- 
2.46.0





This bug report was last modified 145 days ago.

Previous Next


GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.