doc/html/job_state_codes.shtml

<!--#include virtual="header.txt"-->

<h1>Job State Codes</h1>

<p>Each job and job step in the Slurm system has a state assigned to it. How
the state is displayed depends on the method used to identify it.</p>

<h2 id="overview">Overview<a class="slurm_link" href="#overview"></a></h2>

<p>In the Slurm code, there are <b>base states</b> and <b>state flags</b>.
Each job and job step has a base state and may have additional state flags
set. When using
the <a href="rest_quickstart.html">REST API</a>, both the base state and current
flag(s) will be returned.</p>

<p>When the <a href="squeue.html">squeue</a> and <a href="sacct.html">sacct</a>
command report a job state, they represent it as a single state. Both will
recognize all base states but not all state flags. If a recognized flag is
present, it will be reported instead of the base state. Refer to the relevant
command documentation for details.</p>

<p>This page represents all job codes and flags that are represented in the
code. The names provided are the string representations that are used in
user-facing output. For most, the names used in the code are identical, with
<code>JOB_</code> at the start.
For more visibility into the job states and flags, set
<code>DebugFlags=TraceJobs</code> and <code>SlurmctldDebug=verbose</code>
(or higher) in <a href="slurm.conf.html">slurm.conf</a>.</p>

<h2 id="states">Job and step states
<a class="slurm_link" href="#states"></a></h2>

<p>Each job and job step known to the system will have one of the following
states:</p>

<table>
<tbody>
<tr><td><strong>Name</strong></td><td><strong>Description</strong></td></tr>
<tr><td><code>BOOT_FAIL</code></td><td>terminated due to node boot failure</td></tr>
<tr><td><code>CANCELLED</code></td><td>cancelled by user or administrator</td></tr>
<tr><td><code>COMPLETED</code></td><td>completed execution successfully;
	finished with an <a href="job_exit_code.html">exit code</a> of zero on all nodes</td></tr>
<tr><td><code>DEADLINE</code></td><td>terminated due to reaching the latest
	start time that allows the job to reach its deadline given its TimeLimit</td></tr>
<tr><td><code>FAILED</code></td><td>completed execution unsuccessfully;
	non-zero <a href="job_exit_code.html">exit code</a> or other failure condition</td></tr>
<tr><td><code>NODE_FAIL</code></td><td>terminated due to node failure</td></tr>
<tr><td><code>OUT_OF_MEMORY</code></td><td>experienced out of memory error</td></tr>
<tr><td><code>PENDING</code></td><td>queued and waiting for initiation;
	will typically have a <a href="job_reason_codes.html">reason code</a>
	specifying why it has not yet started</td></tr>
<tr><td><code>PREEMPTED</code></td><td>terminated due to
	<a href="preempt.html">preemption</a>; may transition to another state
	based on the configured PreemptMode and job characteristics</td></tr>
<tr><td><code>RUNNING</code></td><td>allocated resources and executing</td></tr>
<tr><td><code>SUSPENDED</code></td><td>allocated resources but execution
	suspended, such as from <a href="preempt.html">preemption</a> or a
	<a href="scontrol.html#OPT_suspend">direct request</a> from an
	authorized user</td></tr>
<tr><td><code>TIMEOUT</code></td><td>terminated due to reaching the time limit,
	such as those configured in <a href="slurm.conf.html">slurm.conf</a> or
	specified for the individual job</td></tr>
</tbody>
</table>

<h2 id="flags">Job and step flags<a class="slurm_link" href="#flags"></a></h2>

<p>Jobs and job steps may have additional flags set:</p>

<table>
<tbody>
<tr><td><strong>Name</strong></td><td><strong>Description</strong></td></tr>
<tr><td><code>COMPLETING</code></td><td>job has finished or been cancelled
	and is performing cleanup tasks, including the
	<a href="prolog_epilog.html">epilog</a> script if present</td></tr>
<tr><td><code>CONFIGURING</code></td><td>job has been allocated nodes and is
	waiting for them to become ready for use, such as booting or rebooting
	the nodes, or running a configured
	<a href="prolog_epilog.html">prolog</a>
	(<code>PrologSlurmctld</code>, or <code>Prolog</code> with
	<code>PrologFlags=Alloc</code> unless <code>NoHold</code> is also
	set); on a job step managed by the
	<a href="slurm.conf.html#OPT_enable_stepmgr">step manager</a>, the step
	has been allocated resources and is waiting to be handed to
	<a href="srun.html">srun</a>, which is normally momentary</td></tr>
<tr><td><code>EXPEDITING</code></td><td>the job is immediately eligible for
	scheduling with the highest possible priority</td></tr>
<tr><td><code>LAUNCH_FAILED</code></td><td>failed to launch on the chosen
	node(s); includes <a href="prolog_epilog.html">prolog</a> failure and
	other failure conditions</td></tr>
<tr><td><code>POWER_UP_NODE</code></td><td>job has been allocated powered down
	nodes and is waiting for them to boot</td></tr>
<tr><td><code>RECONFIG_FAIL</code></td><td>node configuration for job failed</td></tr>
<tr><td><code>REQUEUED</code></td><td>job is being requeued,
	such as from <a href="preempt.html">preemption</a> or a
	<a href="scontrol.html#OPT_requeue">direct request</a> from an
	authorized user</td></tr>
<tr><td><code>REQUEUE_FED</code></td><td>requeued due to conditions of its
	sibling job in a <a href="federation.html">federated</a> setup</td></tr>
<tr><td><code>REQUEUE_HOLD</code></td><td>same as <code>REQUEUED</code> but will
	not be considered for scheduling until it is
	<a href="scontrol.html#OPT_release">released</a></td></tr>
<tr><td><code>RESIZING</code></td><td>the size of the job is changing; prevents
	conflicting job changes from taking place</td></tr>
<tr><td><code>RESV_DEL_HOLD</code></td><td>held due to deleted reservation</td></tr>
<tr><td><code>REVOKED</code></td><td>revoked due to conditions of its sibling
	job in a <a href="federation.html">federated</a> setup</td></tr>
<tr><td><code>SIGNALING</code></td><td>outgoing signal to job is pending</td></tr>
<tr><td><code>SPECIAL_EXIT</code></td><td>same as <code>REQUEUE_HOLD</code> but
	used to identify a <a href="scontrol.html#OPT_State">special situation</a>
	that applies to this job</td></tr>
<tr><td><code>STAGE_OUT</code></td><td>staging out data
	(<a href="burst_buffer.html">burst buffer</a>)</td></tr>
<tr><td><code>STOPPED</code></td><td>received SIGSTOP to suspend the job without
	releasing resources</td></tr>
<tr><td><code>UPDATE_DB</code></td><td>sending an update about the job to the
	database</td></tr>
</tbody>
</table>
<br>

<!--#include virtual="footer.txt"-->