[slurm-users] detailed worker state with sinfo

Guillaume COCHARD guillaume.cochard at cc.in2p3.fr
Mon Jun 27 07:05:29 UTC 2022


You can use `scontrol show nodes` (for all nodes) or `scontrol show node <nodename(s)>` (for one or more nodes). 

For example : 

$ scontrol show node node0001 
NodeName=node0001 Arch=x86_64 

With that it would be quite easy to write a small script if you want a specific output. 


De: "Brian Andrus" <toomuchit at gmail.com> 
À: slurm-users at lists.schedmd.com 
Envoyé: Dimanche 26 Juin 2022 15:36:53 
Objet: Re: [slurm-users] detailed worker state with sinfo 


There are not such differences for draining nodes. 

DRNG = Draining 

DRAIN = Drained 

So if a node is not fully drained yet, it is draining 

It would never be IDLE + DRAINING because it can immediately become drained. 

STATE State of the nodes. Possible states include: allocated, completing, down, drained, draining, fail, failing, future, idle, maint, mixed, perfctrs, planned, power_down, power_up, reserved, and unknown. Their abbreviated forms are: alloc, comp, down, drain, drng, fail, failg, futr, idle, maint, mix, npc, plnd, pow_dn, pow_up, resv, and unk respectively. 

NOTE : The suffix "*" identifies nodes that are presently not responding. 
Brian Andrus 

On 6/26/2022 5:39 AM, [ mailto:z148x at arcor.de | z148x at arcor.de ] wrote: 


if I call "sinfo -o %all", the worker state includes only a single state
word like "DRNG".

It is clearer in the slurm database:
ALLOCATED + DRAIN = DRNG, compared to the db entry.

Where can I get a list of what all the sinfo state names mean compared
to the database?



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.schedmd.com/pipermail/slurm-users/attachments/20220627/4ddf78bd/attachment.htm>

More information about the slurm-users mailing list