Skip to content

Conversation

@karasevb
Copy link
Member

Reseting the ORTE_NODE_FLAG_MAPPED flag after hosts filtering, this
flag is used subsequently and can be affect to the node mapping logic

Signed-off-by: Boris Karasev karasev.b@gmail.com

Reseting the `ORTE_NODE_FLAG_MAPPED` flag after hosts filtering, this
flag is used subsequently and can be affect to the node mapping logic

Signed-off-by: Boris Karasev <karasev.b@gmail.com>
@karasevb
Copy link
Member Author

This PR fixes the nodes ranking failure which is reproduced as described below:

Slurm sbatch file for reproduce the problem:

#!/bin/bash
#SBATCH --job-name=test -w node[1-2]
#SBATCH --nodes=2 --ntasks-per-node=8
#SBATCH --time=01:00:00 --partition=debug

mpirun -n 2 -bind-to core -map-by node -H node1,node2 hostname

output:

[node1:25268] [[52094,0],0] ORTE_ERROR_LOG: Fatal in file base/rmaps_base_ranking.c at line 646
[node1:25268] [[52094,0],0] ORTE_ERROR_LOG: Fatal in file base/odls_base_default_fns.c at line 531

@jsquyres
Copy link
Member

@karasevb Please do not open pull request branches on the main Open MPI repository.

@artpol84
Copy link
Contributor

Sorry about that.
Obviously it was a mistake

@artpol84 artpol84 merged commit 714c8c7 into master Mar 23, 2018
@jsquyres jsquyres deleted the host_filtering branch March 23, 2018 17:28
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

5 participants