<div dir="ltr"><div dir="ltr">This makes sense and I have a unified authorization scheme, but how do the slurm commands (e.g., salloc and squeue) know which slurm head node to talk to? Do I have to run munge on the submitting machines?</div><div dir="ltr"><div><div dir="ltr" class="gmail_signature"><div dir="ltr"><div><br></div>-steve</div></div></div><br></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Sun, Aug 27, 2023 at 5:00 AM <<a href="mailto:slurm-users-request@lists.schedmd.com">slurm-users-request@lists.schedmd.com</a>> wrote:</div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">
Message: 3<br>
Date: Sun, 27 Aug 2023 11:10:18 +0100<br>
From: William Brown <<a href="mailto:william@signalbox.org.uk" target="_blank">william@signalbox.org.uk</a>><br>
To: Slurm User Community List <<a href="mailto:slurm-users@lists.schedmd.com" target="_blank">slurm-users@lists.schedmd.com</a>><br>
Subject: Re: [slurm-users] Submitting jobs from machines outside the<br>
cluster<br>
Message-ID:<br>
<CAGdVoZirihTKz7V=<a href="mailto:yEoVENKM7LcJoE-JrZdaGA%2Bh2wgBNDk57w@mail.gmail.com" target="_blank">yEoVENKM7LcJoE-JrZdaGA+h2wgBNDk57w@mail.gmail.com</a>><br>
Content-Type: text/plain; charset="utf-8"<br>
<br>
The machine that runs the CLI isn't usually a cluster node or run slurmd.<br>
<br>
The control node has to accept the user is who they claim and AFAIK that is<br>
the job of munge. And your onboard and external firewalls must allow the<br>
requisite ports.<br>
<br>
We used Galaxy (see <a href="http://galaxy.eu" rel="noreferrer" target="_blank">galaxy.eu</a> if unfamiliar) and could submit jobs to<br>
various job runners including Slurm. The galaxy node definitely didn't run<br>
any slurm daemons.<br>
<br>
I think you do need a common authentication system between the submitting<br>
node and the cluster, but that may just be what I'm used to.<br>
<br>
<br>
<br>
William Brown<br>
<br>
On Sun, 27 Aug 2023, 07:20 Steven Swanson, <<a href="mailto:sjswanson@ucsd.edu" target="_blank">sjswanson@ucsd.edu</a>> wrote:<br>
<br>
> Can I submit jobs with a computer/docker container that is not part of the<br>
> slurm cluster?<br>
><br>
> I'm trying to set up slurm as the backend for a system with Jupyter<br>
> Notebook-based front end.<br>
><br>
> The jupyter notebooks are running in containers managed by Jupyter Hub,<br>
> which is a mostly turnkey system for providing docker containers that users<br>
> can access via jupyter.<br>
><br>
> I would like the jupyter containers to be able to submit jobs to slurm,<br>
> but making them part of the cluster doesn't seem to make sense because:<br>
><br>
> 1. They are dynamically created and don't have known hostnames.<br>
> 2. They aren't supposed to run jobs.<br>
><br>
> Is there a way to do this? I tried just running slurmd in the jupyter<br>
> containers, but it complained about not being able to figure out its name<br>
> (I think because the container's hostname is not listed in slurm.conf).<br>
><br>
> My fall back solution is to use ssh to connect to the slurm head node and<br>
> run jobs there, but that seems kludgy.<br>
><br>
> -steve<br>
><br>
-------------- next part --------------<br>
An HTML attachment was scrubbed...<br>
URL: <<a href="https://urldefense.com/v3/__http://lists.schedmd.com/pipermail/slurm-users/attachments/20230827/fa37f4ae/attachment-0001.htm__;!!Mih3wA!H6jtRH1Co7St0JJ5HiauL4ay4-Eh6LH0U82KsCeTJHkkjp-ljokZ6CgLmcjQ0N_QpvgxDYbrcQv4bQYRUvStqM0vuZcLCGlWVg$" rel="noreferrer" target="_blank">https://urldefense.com/v3/__http://lists.schedmd.com/pipermail/slurm-users/attachments/20230827/fa37f4ae/attachment-0001.htm__;!!Mih3wA!H6jtRH1Co7St0JJ5HiauL4ay4-Eh6LH0U82KsCeTJHkkjp-ljokZ6CgLmcjQ0N_QpvgxDYbrcQv4bQYRUvStqM0vuZcLCGlWVg$</a> ><br>
<br>
End of slurm-users Digest, Vol 70, Issue 36<br>
*******************************************<br>
</blockquote></div></div>