Why would you run a slurmctld on a *submit* host? You only need the controller daemon on, well, the controllers (what I would still call 'queue masters' :) ). Personally I'd make quite sure that no-one apart from admins has rights to log in to those, really!
In fact, you don't need to run any daemon on a submit host; it just needs access to the binaries and the cluster config. We run Centos, so I build the rpms; all I install on the login node(s) is 'slurm' - neither 'slurmd' nor 'slurmctld' are required. Which is certainly not something you couldn't install on a bunch or workstations, especially if they are managed machines. Tina On 26/11/2018 23:23, Goetz, Patrick G wrote: > I'm a little confused about how this would work. For example, where > does slurmctld run? And if on each submit host, why aren't the control > daemons stepping all over each other? > > On 11/22/18 6:38 AM, Stu Midgley wrote: >> indeed. >> >> All our workstations are submit hosts and in the queue, so people can >> run jobs on their local host if they want. >> >> We have a GUI tightly integrated with our environment for our staff to >> submit and monitor their jobs from (they don't have to touch a single >> job script). >> >> On Thu, Nov 22, 2018 at 6:28 PM Tina Friedrich >> <tina.friedr...@it.ox.ac.uk <mailto:tina.friedr...@it.ox.ac.uk>> wrote: >> >> I really don't want to start a flaming discussion on this - but I don't >> think it's an unusual situation. I have, in likewise roughtly 15 years >> of doing this, not ever worked anywhere where people didn't have a GUI >> to submit from. It's always been a case of 'Wand to use the cluster? >> We'll make your workstation a submit host.' >> >> I think it's a pretty standard way of handling things it you are an >> institute that runs their own (maybe small) cluster, especially if the >> workstations are also managed machine. >> >> Tina >> >> On 21/11/2018 23:26, Christopher Samuel wrote: >> > On 22/11/18 5:04 am, Mahmood Naderan wrote: >> > >> >> The idea is to have a job manager that find the best node for a >> newly >> >> submitted job. If the user has to manually ssh to a node, why one >> >> should use slurm or any other thing? >> > >> > You are in a really really unusual situation - in 15 years I've >> not come >> > across a situation before this where a user would have GUI access >> to a >> > system that can submit jobs directly to a cluster like you can. >> > >> > I'm not sure why Slurm has this restriction but it might be that >> you can >> > start up an xterm, change your $DISPLAY to be localhost:0 and see >> if you >> > can start an X11 application from that. It might be that you'll >> need to >> > add an xauth cookie for localhost to get that going. >> > >> > If it does work then (hopefully) you can use that trick to fire >> up jobs >> > with X11 display forwarding. >> > >> > All the best, >> > Chris >> >> >> >> -- >> Dr Stuart Midgley >> sdm...@gmail.com <mailto:sdm...@gmail.com>
