12552: Improve PrioritySpread docs.

[arvados.git] / doc / install / crunch2-slurm / install-dispatch.html.textile.liquid

diff --git a/doc/install/crunch2-slurm/install-dispatch.html.textile.liquid b/doc/install/crunch2-slurm/install-dispatch.html.textile.liquid
index d01326a15a3f26e0edde8f9f773f24498a079b88..9784266e68b4df509e0ce55f3db8118bbd77d99e 100644 (file)
--- a/doc/install/crunch2-slurm/install-dispatch.html.textile.liquid
+++ b/doc/install/crunch2-slurm/install-dispatch.html.textile.liquid
@@ -4,6 +4,11 @@ navsection: installguide
 title: Install the SLURM dispatcher
 
 ...
+{% comment %}
+Copyright (C) The Arvados Authors. All rights reserved.
+
+SPDX-License-Identifier: CC-BY-SA-3.0
+{% endcomment %}
 
 The SLURM dispatcher can run on any node that can submit requests to both the Arvados API server and the SLURM controller.  It is not resource-intensive, so you can run it on the API server node.
 
@@ -28,7 +33,7 @@ On Debian-based systems:
 
 h2. Create a dispatcher token
 
-Create a privileged Arvados API token for use by the dispatcher. If you have multiple dispatch processes, you should give each one a different token.  *On the API server*, run:
+Create an Arvados superuser token for use by the dispatcher. If you have multiple dispatch processes, you should give each one a different token.  *On the API server*, run:
 
 <notextile>
 <pre><code>apiserver:~$ <span class="userinput">cd /var/www/arvados-api/current</span>
@@ -64,8 +69,8 @@ Override Keep service discovery with a predefined list of Keep URIs. This can be
 
 <notextile>
 <pre><code class="userinput">Client:
-  APIHost: <b>zzzzz.arvadosapi.com</b>
-  AuthToken: <b>zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz</b>
+  APIHost: zzzzz.arvadosapi.com
+  AuthToken: zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz
   KeepServiceURIs:
   - <b>http://127.0.0.1:25107</b>
 </code></pre>
@@ -76,22 +81,32 @@ h3. PollPeriod
 crunch-dispatch-slurm polls the API server periodically for new containers to run.  The @PollPeriod@ option controls how often this poll happens.  Set this to a string of numbers suffixed with one of the time units @ns@, @us@, @ms@, @s@, @m@, or @h@.  For example:
 
 <notextile>
-<pre><code class="userinput">Client:
-  APIHost: <b>zzzzz.arvadosapi.com</b>
-  AuthToken: <b>zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz</b>
-PollPeriod: <b>3m30s</b>
+<pre><code class="userinput">PollPeriod: <b>3m30s</b>
+</code></pre>
+</notextile>
+
+h3. PrioritySpread
+
+crunch-dispatch-slurm adjusts the "nice" values of its SLURM jobs to ensure containers are prioritized correctly relative to one another. This option tunes the adjustment mechanism.
+* If non-Arvados jobs run on your SLURM cluster, and your Arvados containers are waiting too long in the SLURM queue because their "nice" values are too high for them to compete with other SLURM jobs, you should use a smaller PrioritySpread value.
+* If you have an older SLURM system that limits nice values to 10000, a smaller @PrioritySpread@ can help avoid reaching that limit.
+* In other cases, a larger value is beneficial because it reduces the total number of adjustments made by executing @scontrol@.
+
+The smallest usable value is @1@. The default value of @10@ is used if this option is zero or negative. Example:
+
+<notextile>
+<pre><code class="userinput">PrioritySpread: <b>1000</b>
 </code></pre>
 </notextile>
 
+
+
 h3. SbatchArguments
 
 When crunch-dispatch-slurm invokes @sbatch@, you can add switches to the command by specifying @SbatchArguments@.  You can use this to send the jobs to specific cluster partitions or add resource requests.  Set @SbatchArguments@ to an array of strings.  For example:
 
 <notextile>
-<pre><code class="userinput">Client:
-  APIHost: <b>zzzzz.arvadosapi.com</b>
-  AuthToken: <b>zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz</b>
-SbatchArguments:
+<pre><code class="userinput">SbatchArguments:
 - <b>"--partition=PartitionName"</b>
 </code></pre>
 </notextile>
@@ -101,10 +116,7 @@ h3. CrunchRunCommand: Dispatch to SLURM cgroups
 If your SLURM cluster uses the @task/cgroup@ TaskPlugin, you can configure Crunch's Docker containers to be dispatched inside SLURM's cgroups.  This provides consistent enforcement of resource constraints.  To do this, use a crunch-dispatch-slurm configuration like the following:
 
 <notextile>
-<pre><code class="userinput">Client:
-  APIHost: <b>zzzzz.arvadosapi.com</b>
-  AuthToken: <b>zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz</b>
-CrunchRunCommand:
+<pre><code class="userinput">CrunchRunCommand:
 - <b>crunch-run</b>
 - <b>"-cgroup-parent-subsystem=memory"</b>
 </code></pre>
@@ -120,15 +132,24 @@ You can work around this issue by disabling the Docker daemon's systemd integrat
 
 {% include 'notebox_end' %}
 
+h3. CrunchRunCommand: Using host networking for containers
+
+Older Linux kernels (prior to 3.18) have bugs in network namespace handling which can lead to compute node lockups.  This by is indicated by blocked kernel tasks in "Workqueue: netns cleanup_net".   If you are experiencing this problem, as a workaround you can disable use of network namespaces by Docker across the cluster.  Be aware this reduces container isolation, which may be a security risk.
+
+<notextile>
+<pre><code class="userinput">CrunchRunCommand:
+- <b>crunch-run</b>
+- <b>"-container-enable-networking=always"</b>
+- <b>"-container-network-mode=host"</b>
+</code></pre>
+</notextile>
+
 h3. MinRetryPeriod: Rate-limit repeated attempts to start containers
 
 If SLURM is unable to run a container, the dispatcher will submit it again after the next PollPeriod. If PollPeriod is very short, this can be excessive. If MinRetryPeriod is set, the dispatcher will avoid submitting the same container to SLURM more than once in the given time span.
 
 <notextile>
-<pre><code class="userinput">Client:
-  APIHost: <b>zzzzz.arvadosapi.com</b>
-  AuthToken: <b>zzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzzz</b>
-MinRetryPeriod: <b>30s</b>
+<pre><code class="userinput">MinRetryPeriod: <b>30s</b>
 </code></pre>
 </notextile>