[slurm-users] NIC gres types and lack of device files?

[Chris Samuel]

Hi folks,

Just curious if anyone out there is currently using the "nic" gres type and 
how do you deal with the fact that the Files= entry in gres.conf has to refer 
to an existing device when network devices generally don't appear under /dev?

Do you just hack with with "touch" or "mkdev" ?

The bug that introduced the code was:

https://bugs.schedmd.com/show_bug.cgi?id=1718#c3

and the example given was:

Name=nic File=/dev/mlx4_0 CPUs=0-1
Name=nic File=/dev/mlx4_1 CPUs=2-3

but nothing like those names will exist with Mellanox OFED.

[root at spartan-gpgpu009 ~]# ibdev2netdev
mlx5_0 port 1 ==> p1p1 (Up)
mlx5_1 port 1 ==> p1p2 (Up)
mlx5_2 port 1 ==> p2p1 (Down)
mlx5_3 port 1 ==> p2p2 (Down)
[root at spartan-gpgpu009 ~]# find /dev/ -name mlx*
[root at spartan-gpgpu009 ~]# find /dev/ -name p1*
[root at spartan-gpgpu009 ~]#


cheers,
Chris
-- 
 Christopher Samuel        Senior Systems Administrator
 Melbourne Bioinformatics - The University of Melbourne
 Email: samuel at unimelb.edu.au Phone: +61 (0)3 903 55545