Discussion:
[OMPI users] Unable to open a shared object libsmartio-rdmav17.so
Llolsten Kaonga
2018-08-24 14:23:46 UTC
Permalink
Hello all,



This may be a silly question but I hope that someone does know the answer.



We use Open MPI to run the Intel Benchmarks to test InfiniBand and RoCE
network fabrics. We recently installed OFED-4.17 and when we attempt to run
the tests, we see the error below.



Command:

/usr/local/bin/mpirun --allow-run-as-root --mca btl openib,self,vader --mca
pml ob1 -np 8 -hostfile /root/mpi-hosts /usr/local/bin/IMB-MPI1



Result:

libibverbs: Warning: couldn't load driver 'libsmartio-rdmav17.so':
libsmartio-rdmav17.so: cannot open shared object file: No such file or
directory

libibverbs: Warning: couldn't load driver 'libsmartio-rdmav17.so':
libsmartio-rdmav17.so: cannot open shared object file: No such file or
directory

libibverbs: Warning: couldn't load driver 'libsmartio-rdmav17.so':
libsmartio-rdmav17.so: cannot open shared object file: No such file or
directory

libibverbs: Warning: couldn't load driver 'libsmartio-rdmav17.so':
libsmartio-rdmav17.so: cannot open shared object file: No such file or
directory

libibverbs: Warning: couldn't load driver 'libsmartio-rdmav17.so':
libsmartio-rdmav17.so: cannot open shared object file: No such file or
directory

[sm-node-02][[44319,1],6][btl_openib_component.c:1670:init_one_device] error
obtaining device attributes for mlx5_0 errno says Success

[sm-node-02][[44319,1],5][btl_openib_component.c:1670:init_one_device] error
obtaining device attributes for mlx5_0 errno says Success

[sm-node-02][[44319,1],4][btl_openib_component.c:1670:init_one_device] error
obtaining device attributes for mlx5_0 errno says Success



The folks who build OFED believe libsmartio-rdmav17.so is not part of the
OFED package. It is not in RDMA-Core. I have searched for information on
this object and can't seem to find anything. If anyone knows anything about
it or (especially) thinks that we should change our mpirun command options,
or has pointers to where I should direct this question, I would appreciate
the help.



OS: CentOS 7.4;

(kernel: 4.17.14-1.el7.elrepo.x86_64)

OFED: OFED-4.17-20180822-1352
(https://www.openfabrics.org/downloads/OFED/ofed-4.17-daily/OFED-4.17-201808
22-1352.tgz)



I will be happy to provide any additional information if needed.



Thanks.

--

Llolsten
Jeff Squyres (jsquyres) via users
2018-08-25 01:15:59 UTC
Permalink
I'm afraid the error message you're getting is from libibverbs; it's trying to load a plugin named libsmartio-rdmav17.so. That's not part of Open MPI, sorry.

That likely means that some dependency of libsmartio-rdmav17.so wasn't found, and the run-time loading of the plugin failed (vs. not being able to find the libsmartio-rdmav17.so file).

You might want to track down where you got the libsmartio-rdmav17.so file from.
Post by Llolsten Kaonga
Hello all,
This may be a silly question but I hope that someone does know the answer.
We use Open MPI to run the Intel Benchmarks to test InfiniBand and RoCE network fabrics. We recently installed OFED-4.17 and when we attempt to run the tests, we see the error below.
/usr/local/bin/mpirun --allow-run-as-root --mca btl openib,self,vader --mca pml ob1 -np 8 -hostfile /root/mpi-hosts /usr/local/bin/IMB-MPI1
libibverbs: Warning: couldn't load driver 'libsmartio-rdmav17.so': libsmartio-rdmav17.so: cannot open shared object file: No such file or directory
libibverbs: Warning: couldn't load driver 'libsmartio-rdmav17.so': libsmartio-rdmav17.so: cannot open shared object file: No such file or directory
libibverbs: Warning: couldn't load driver 'libsmartio-rdmav17.so': libsmartio-rdmav17.so: cannot open shared object file: No such file or directory
libibverbs: Warning: couldn't load driver 'libsmartio-rdmav17.so': libsmartio-rdmav17.so: cannot open shared object file: No such file or directory
libibverbs: Warning: couldn't load driver 'libsmartio-rdmav17.so': libsmartio-rdmav17.so: cannot open shared object file: No such file or directory
[sm-node-02][[44319,1],6][btl_openib_component.c:1670:init_one_device] error obtaining device attributes for mlx5_0 errno says Success
[sm-node-02][[44319,1],5][btl_openib_component.c:1670:init_one_device] error obtaining device attributes for mlx5_0 errno says Success
[sm-node-02][[44319,1],4][btl_openib_component.c:1670:init_one_device] error obtaining device attributes for mlx5_0 errno says Success
The folks who build OFED believe libsmartio-rdmav17.so is not part of the OFED package. It is not in RDMA-Core. I have searched for information on this object and can’t seem to find anything. If anyone knows anything about it or (especially) thinks that we should change our mpirun command options, or has pointers to where I should direct this question, I would appreciate the help.
OS: CentOS 7.4;
(kernel: 4.17.14-1.el7.elrepo.x86_64)
OFED: OFED-4.17-20180822-1352 (https://www.openfabrics.org/downloads/OFED/ofed-4.17-daily/OFED-4.17-20180822-1352.tgz)
I will be happy to provide any additional information if needed.
Thanks.
--
Llolsten
_______________________________________________
users mailing list
https://lists.open-mpi.org/mailman/listinfo/users
--
Jeff Squyres
***@cisco.com

Loading...