blob: d94df5bdd0b48c54764a01097cbe2b0e762df2c1 [file] [log] [blame]
<!--#include virtual="header.txt"-->
<h1>REST API Quick Start Guide</h1>
<p>Slurm provides a <a href="rest.html">REST API</a> through the slurmrestd
daemon, using <a href="jwt.html">JSON Web Tokens</a> for authentication.
This page provides a brief tutorial for setting up these components.</p>
<p>
See also:
<ul>
<li><a href="rest.html">REST API Reference</a></li>
<li><a href="rest_api.html">REST API Implementation Details</a></li>
<li><a href="slurmrestd.html">slurmrestd man page</a></li>
</ul>
</p>
<h2 id="contents">Contents<a class="slurm_link" href="#contents"></a></h2>
<ul>
<li><a href="#prereq">Prerequisites</a></li>
<li><a href="#quick_start">Quick Start</a>
<ul>
<li><a href="#systemd">Running with systemd</a></li>
<li><a href="#customization">Customizing slurmrestd.service</a></li>
</ul>
</li>
<li><a href="#basic_usage">Basic Usage</a>
<ul>
<li><a href="#tokens">Token management</a></li>
</ul>
</li>
<li><a href="#common_issues">Common Issues</a>
<ul>
<li><a href="#bind_socket">Unable to bind socket</a></li>
<li><a href="#connection_refused">Connection refused</a></li>
<li><a href="#authentication_error">Protocol authentication error (HTTP 500)
</a></li>
<li><a href="#invalid_url">Unable to find requested URL (HTTP 404)</a></li>
<li><a href="#rejected_token">Rejecting thread config token (HTTP 401)
</a></li>
<li><a href="#unexpected_character">Unexpected URL character (HTTP 400)
</a></li>
<li><a href="#slurm_commands">Other Slurm commands not working</a></li>
</ul>
</li>
</ul>
<h2 id="prereq">Prerequisites
<a class="slurm_link" href="#prereq"></a></h2>
<p>The following development libraries are required at compile time
in order for slurmrestd to be compiled:</p>
<ul>
<li><a href="download.html#httpparser">HTTP Parser</a> (&gt;= v2.6.0)
</li>
<li><a href="download.html#yaml">LibYAML</a> (optional, &gt;= v0.2.5)
</li>
<li><a href="download.html#json">JSON-C</a> (&gt;= v1.12.0)</li>
<li><a href="download.html#jwt">JWT Authentication</a> (optional)</li>
</ul>
<h2 id="quick_start">Quick Start
<a class="slurm_link" href="#quick_start"></a></h2>
<p>This may be done on a dedicated REST API machine or
your existing 'slurmctld' machine, depending on demand.</p>
<ol>
<li>Install components for slurmrestd
<ul>
<li>DEB: <code>slurm-smd slurm-smd-slurmrestd</code></li>
<li>RPM: <code>slurm slurm-slurmrestd</code> (requires
<code>--with slurmrestd</code> at build time)</li>
</ul></li>
<li>Set up <a href="jwt.html">JSON Web Tokens</a> for authentication</li>
<li>Ensure <code>/etc/slurm/slurm.conf</code> is present and correct for your
cluster (see <a href="quickstart_admin.html">Quick Start Admin Guide</a> and
<a href="slurm.conf.html">slurm.conf man page</a>)</li>
<li>Run <b>slurmrestd</b> (see <a href="#systemd">below</a> for systemd
instructions) on your preferred [HOST]:PORT combination
(':6820' is the default for production)
<pre>export SLURM_JWT=daemon
export SLURMRESTD_DEBUG=debug
slurmrestd &lt;host&gt;:&lt;port&gt;
</pre>
Adjust SLURMRESTD_DEBUG to the desired level of output (as described on the
<a href="slurmrestd.html#OPT_SLURMRESTD_DEBUG">man page</a>)</li>
</ol>
<h3 id="systemd">Running with systemd
<a class="slurm_link" href="#systemd"></a></h3>
Slurm ships with a <b>slurmrestd</b> service for systemd,
however, it might require some additional setup to run properly.
This section assumes you have either installed slurmrestd using DEB/RPM packages
or built it manually such that the files are in the same places.
<ol>
<li>Select a local user account to use to run the slurmrestd
(<b>not</b> root or SlurmUser to prevent privilege escalation)</li>
<li>Configure slurmrestd service to use this user and associated group.
This can be accomplished in either of two ways:
<ul>
<li>Edit <code>/etc/default/slurmrestd</code>
or <code>/etc/sysconfig/slurmrestd</code>
<br>Add <code>-u USERNAME</code> and <code>-g GROUPNAME</code>
to <code>SLURMRESTD_OPTIONS</code></li>
<li>Run <code>systemctl edit slurmrestd</code> to edit overrides for the
service.
<br>Add content like this to the prescribed location in the overrides file:
<pre>
[Service]
User=USERNAME
Group=GROUPNAME
</pre></li>
</ul>
</li>
<li>Configure the socket for slurmrestd. This may be accomplished by
creating/changing permissions on the parent directory and/or changing the path
to the socket in the service file.
<ul>
<li><b>Permissions</b>: The user running the service must have write+execute
permissions on directory that will contain the socket</li>
<li><b>Changing socket</b>: Currently the way to change or disable the socket
is to modify the 'ExecStart' line of the service (this is known to be an
inelegant procedure and will be changed in the next major release)
<ol>
<li>Run <code>systemctl edit slurmrestd</code></li>
<li>Add the following contents to the <code>[Service]</code> section:
<pre>
ExecStart=
ExecStart=/usr/sbin/slurmrestd $SLURMRESTD_OPTIONS
Environment=SLURMRESTD_LISTEN=:6820
</pre></li>
<li>Adjust the assignment of SLURMRESTD_LISTEN to contain the socket(s) you want
the daemon to listen on.</li>
</ol>
<p><b>NOTE</b>: Currently the way to change or disable the socket is to modify
the 'ExecStart' line of the service. This is known to be an inelegant procedure
and will be changed in the next major release (after 23.11).</p>
</li>
</ul>
</li>
</ol>
<h3 id="customization">Customizing slurmrestd.service
<a class="slurm_link" href="#customization"></a></h3>
<p>The next major release (after 23.11) changes the operation of
the default service file and may break existing overrides. If you have
overridden <code>ExecStart=</code> to contain any TCP/UNIX sockets directly, it
will cause the service to fail if it duplicates any sockets contained in
SLURMRESTD_LISTEN. These overrides will need to be changed after upgrading.</p>
<p>The default <code>slurmrestd.service</code> file has two intended ways of
customizing its operation:</p>
<ol>
<li><b>Environment files</b>:
The service will read environment variables from two files:
<code>/etc/default/slurmrestd</code> and <code>/etc/sysconfig/slurmrestd</code>.
You may set any environment variables recognized by
<a href="slurmrestd.html#SECTION_ENVIRONMENT-VARIABLES">slurmrestd</a>,
but the following are particularly relevant:
<ul>
<li>SLURMRESTD_OPTIONS: CLI options to add to the slurmrestd command (see
<a href="slurmrestd.html">slurmrestd</a>)</li>
<li>SLURMRESTD_LISTEN: Comma-delimited list of host:port pairs or unix sockets
to listen on
<br><b>NOTE</b>: if this duplicates what is already set in the
'ExecStart' line in the service file, it will fail</li>
</ul>
</li>
<li><b>Service editing</b>: Systemd has a built in way to edit services
by running <code>systemctl edit slurmrestd</code>.
<ul>
<li>This will create an override file in '/etc/systemd/' containing directives
that will add to or replace directives in the default unit in '/lib/systemd/'.
</li>
<li>Note that the override file must have the appropriate section declaration(s)
for the directives you use (e.g., <code>[Service]</code>).</li>
<li>Changes may be reverted by running <code>systemctl revert slurmrestd</code>
</li>
</ul>
</li>
</ol>
<h2 id="basic_usage">Basic Usage
<a class="slurm_link" href="#basic_usage"></a></h2>
<ol>
<li>Find the latest supported API version
<pre>slurmrestd -d list</pre></li>
<li>Get an authentication token for JWT
<pre>unset SLURM_JWT; export $(scontrol token)</pre>
<ul>
<li>This ensures an old token doesn't prevent a new one from being issued</li>
<li>By default, tokens will expire after 1800 seconds (30 minutes).
Add <code>lifespan=SECONDS</code> to the 'scontrol' command to change this.</li>
</ul>
</li>
<li>Run a basic curl command to hit the API when listening on a TCP host:port
<pre>curl -s -o "/tmp/curl.log" -k -vvvv \
-H X-SLURM-USER-TOKEN:$SLURM_JWT \
-X GET 'http://&lt;server&gt;:&lt;port&gt;/slurm/v0.0.&lt;api-version&gt;/diag'
</pre>
<ul>
<li>Replace the <b>server</b>, <b>port</b>, and <b>api-version</b>
with the appropriate values.</li>
<li>Examine the output to ensure the response was <b>200 OK</b>,
and examine <b>/tmp/curl.log</b> for a valid JSON response.</li>
<li>Try other endpoints described in the <a href="rest_api.html">API Reference
</a>. Change <b>GET</b> to the correct method for the endpoint.</li>
</ul>
</li>
<li>Alternate command to use the UNIX socket instead
<pre>
curl -s -o "/tmp/curl.log" -k -vvvv \
-H X-SLURM-USER-TOKEN:$SLURM_JWT \
--unix-socket /path/to/slurmrestd.socket \
'http://&lt;server&gt;/slurm/v0.0.&lt;api-version&gt;/diag'
</pre>
<ul>
<li>Replace the <b>path</b>, <b>server</b>, and <b>api-version</b>
with the appropriate values.</li>
<li>Examine the output to ensure the response was <b>200 OK</b>,
and examine <b>/tmp/curl.log</b> for a valid JSON response.</li>
</ul>
</li>
</ol>
<h3 id="tokens">Token management
<a class="slurm_link" href="#tokens"></a></h3>
<p>This guide provides a simple overview using <code>scontrol</code> to
obtain tokens. This is a basic introductory approach that in many cases
should be disabled in favor of more sophisticated token management.
Refer to the <a href="jwt.html">JWT page</a> for more details.</p>
<h2 id="common_issues">Common Issues
<a class="slurm_link" href="#common_issues"></a></h2>
<p>In general, look out for these things:</p>
<ol>
<li>Validity of authentication token in <code>SLURM_JWT</code></li>
<li>Hostname and port number</li>
<li>API version and endpoint</li>
<li>Log output of slurmrestd</li>
</ol>
<h3 id="bind_socket">Unable to bind socket
<a class="slurm_link" href="#bind_socket"></a></h3>
<p>This may be due to a permissions issue while attempting to set up the socket.
Check the log output from slurmrestd for the path to the socket.
Ensure that the user running the slurmrestd service has permissions to the
parent directory of the configured socket path, or change/remove the socket path
as <a href="#systemd">described above</a>.</p>
<p>If it says "<b>Address already in use</b>", check the command being run
and the contents of "SLURMRESTD_LISTEN" for duplicates of the same TCP or UNIX
socket.</p>
<h3 id="connection_refused">Connection refused
<a class="slurm_link" href="#connection_refused"></a></h3>
<p>Verify that slurmrestd is running and listening on the port you are
attempting to connect to.</p>
<h3 id="authentication_error">Protocol authentication error (HTTP 500)
<a class="slurm_link" href="#authentication_error"></a></h3>
<p>One common authentication problem is an expired token. Request a new one:
<pre>unset SLURM_JWT; export $(scontrol token)</pre></p>
<p>This solution also applies to an HTTP 401 error caused by no authentication
token being sent at all. This may appear in the slurmrestd logs as
"Authentication does not apply to request."</p>
<p>Otherwise, consult the logs on the <b>slurmctld</b> and <b>slurmdbd</b>.</p>
<h3 id="invalid_url">Unable to find requested URL (HTTP 404)
<a class="slurm_link" href="#invalid_url"></a></h3>
<p>Check the <a href="rest_api.html">API Reference</a> page to ensure you're
using a valid URL and the correct method for it. Pay attention to the path as
there are different endpoints for <b>slurm</b> and <b>slurmdbd</b>.</p>
<h3 id="rejected_token">Rejecting thread config token (HTTP 401)
<a class="slurm_link" href="#rejected_token"></a></h3>
<p>Check that slurmrestd has loaded the <b>auth/jwt</b> plugin.
You should see a debug message like this:
<pre>slurmrestd: debug: auth/jwt: init: JWT authentication plugin loaded</pre>
If it didn't load jwt, run this in the terminal you're using for slurmrestd:
<pre>export SLURM_JWT=daemon</pre>
</p>
<h3 id="unexpected_character">Unexpected URL character (HTTP 400)
<a class="slurm_link" href="#unexpected_character"></a></h3>
<p>Check the request URL and slurmrestd logs for characters that may be causing
the URL to be parsed incorrectly. Use the appropriate URL encoding sequence in
place of the problematic character (e.g., <b>/</b> = <b>%2F</b>).
<pre>
... -X GET "localhost:8080/slurmdb/v0.0.40/jobs?submit_time=02/28/24"
### 400 BAD REQUEST
... -X GET "localhost:8080/slurmdb/v0.0.40/jobs?submit_time=02%2F28%2F24"
### 200 OK
</pre>
</p>
<h3 id="slurm_commands">Other slurm commands not working
<a class="slurm_link" href="#slurm_commands"></a></h3>
<p>If SLURM_JWT is set, other slurm commands will attempt to use JWT
authentication, causing failures. This can be fixed by clearing the variable:
<pre>unset SLURM_JWT</pre></p>
<p style="text-align:center;">Last modified 29 February 2024</p>
<!--#include virtual="footer.txt"-->