<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body text="#000000" bgcolor="#FFFFFF">
<p>I recommend the LLN option for partitions:</p>
<dl compact="compact">
<dt><b>LLN</b></dt>
<dd>
Schedule resources to jobs on the least loaded nodes (based upon
the number
of idle CPUs). This is generally only recommended for an
environment with
serial jobs as idle resources will tend to be highly fragmented,
resulting
in parallel jobs being distributed across many nodes.
Note that node <b>Weight</b> takes precedence over how many
idle resources are
on each node.
Also see the <b>SelectParameters</b> configuration parameter <b>CR_LLN</b>
to
use the least loaded nodes in every partition.
</dd>
</dl>
<p>-Paul Edmon-<br>
</p>
<div class="moz-cite-prefix">On 11/15/2018 4:25 AM, Aravindh
Sampathkumar wrote:<br>
</div>
<blockquote type="cite"
cite="mid:1542273900.388727.1577726464.120DC159@webmail.messagingengine.com">
<meta http-equiv="content-type" content="text/html; charset=UTF-8">
<title></title>
<style type="text/css">p.MsoNormal,p.MsoNoSpacing{margin:0}</style>
<div style="font-size:16px;">Hi All.<br>
</div>
<div style="font-size:16px;"><br>
</div>
<div style="font-size:16px;">I'm having some trouble finding
appropriate section of the documentation to change slurm
resource allocation policy. <br>
</div>
<div style="font-size:16px;"><br>
</div>
<div style="font-size:16px;">We have configured CPU and memory as
consumable resources, and our nodes can run multiple jobs as
long as there are CPU memory available. <br>
</div>
<div style="font-size:16px;"><br>
</div>
<div style="font-size:16px;">What I want is for Slurm to spread
jobs across all available servers in a partition instead of
loading up few servers while others are idling. <br>
</div>
<div style="font-size:16px;"><br>
</div>
<div style="font-size:16px;">For example, I have a partition nav
which has 5 compute nodes(node[1-5]) dedicated to it. <br>
</div>
<div style="font-size:16px;">when users submit 3 jobs to nav
partition, each requesting 1 CPU core and 1 GB of memory, SLURM
schedules all the jobs in node1 because it has enough CPU cores
and memory to satisfy job requirements. nodes - 2,3,4,5 are
idle. <br>
</div>
<div style="font-size:16px;"><br>
</div>
<div style="font-size:16px;">What I want instead is for slurm to
schedule job1 to node1, job2 to node2, job3 to node3.. and then
in the future if there are more jobs than there are nodes, slurm
must utilise the rest of resources available in node1.<br>
</div>
<div style="font-size:16px;"><br>
</div>
<div style="font-size:16px;"><br>
</div>
<div style="font-size:16px;">Why? <br>
</div>
<div style="font-size:16px;">A small group that is using this
partition is concerned that all their jobs get scheduled on the
same node, and they need to share network bandwidth, and
bandwidth to local disk. If they were spread out instead, they
could use better bandwidth. <br>
</div>
<div style="font-size:16px;"><br>
</div>
<div style="font-size:16px;">Appreciate any advice how I can make
this happen. <br>
</div>
<div style="font-size:16px;"><br>
</div>
<div style="font-size:16px;">Thanks,<br>
</div>
<div id="sig56753105">
<div class="signature"> Aravindh Sampathkumar<br>
</div>
<div class="signature"> <a class="moz-txt-link-abbreviated" href="mailto:aravindh@fastmail.com">aravindh@fastmail.com</a><br>
</div>
<div class="signature"><br>
</div>
</div>
<div style="font-size:16px;"><br>
</div>
</blockquote>
</body>
</html>