I use the Feature field in node specifications to accomplish this. From
man slurm.conf(5):
Feature
A comma delimited list of arbitrary strings indicative of
some characteristic associated with the node. There is no value associated
with a feature at this time, a node either has a feature or
it does not. If desired a feature may contain a numeric component
indicat‐
ing, for example, processor speed. By default a node has no
features. Also see Gres.
You then specify --constraint=... to select nodes based on features you
want.
I'm not sure about a Slurm bible, but LLNL does have some additional
documentation:
https://hpc.llnl.gov/banks-jobs/running-jobs/slurm-user-manual
On Thu, Jun 1, 2017 at 8:46 AM Belgin, Mehmet <[email protected]>
wrote:
> Hello everyone,
>
> I was wondering what would be the best way to configure each node with
> custom attributes. For example “ib” for IB nodes, ’ssd’ for SSD local
> disks, you get the point. I’m hoping it’s possible to make up any attribute
> name without being limited to a specific list.
>
> I’d also like to ask where can I find a ’slurm bible’. The documents and
> tutorials on the SchedMD do an excellent job getting slurm up and running
> in no time, but I have a feeling that they don’t include all of the details
> for more advanced configurations.
>
> Thanks in advance!
> -Mehmet
>
>
>
>
>
