openmpi 1.10.2-8ubuntu1 fails running job with cpuset started from torque 5
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
openmpi (Ubuntu) |
New
|
Undecided
|
Unassigned |
Bug Description
OpenMPI 1.10.2 has a bug handling cpusets. Here is a link to the mailing list discussion including the patch I'll attach:
https:/
I encountered this error by adding torque 5.1.1.2 to the system including cpuset support. When I run a job I get this error:
-------
A request for multiple cpus-per-proc was given, but a directive
was also give to map to an object level that has less cpus than
requested ones:
#cpus-per-proc: 1
number of cpus: 0
map-by: BYSOCKET
Please specify a mapping level that has more cpus, or else let us
define a default mapping that will allow multiple cpus-per-proc.
-------
Adding the patch to the deb package and rebuilding it resolved the issue.
ProblemType: Bug
DistroRelease: Ubuntu 16.04
Package: openmpi-bin 1.10.2-8ubuntu1
ProcVersionSign
Uname: Linux 4.4.0-21-generic x86_64
ApportVersion: 2.20.1-0ubuntu2
Architecture: amd64
CurrentDesktop: Unity
Date: Tue May 3 12:11:00 2016
InstallationDate: Installed on 2016-04-27 (6 days ago)
InstallationMedia: Ubuntu 16.04 LTS "Xenial Xerus" - Release amd64 (20160420.1)
SourcePackage: openmpi
UpgradeStatus: No upgrade log present (probably fresh install)
The attachment "fixes hwloc cpuset support" seems to be a patch. If it isn't, please remove the "patch" flag from the attachment, remove the "patch" tag, and if you are a member of the ~ubuntu-reviewers, unsubscribe the team.
[This is an automated message performed by a Launchpad user owned by ~brian-murray, for any issues please contact him.]