Difference between revisions of "Getting started on clusters"

From Earlham CS Department
Jump to navigation Jump to search
Line 42: Line 42:
  
 
Slurm is our batch scheduler.
 
Slurm is our batch scheduler.
 +
 +
You can check that it's working by running: <code>srun -l hostname</code>
 +
 +
You can submit a job in a script with the following: <code>sbatch my_good_script.sbatch</code>
 +
 +
Here's an example of a batch file:
 +
<pre>
 +
#!/bin/sh
 +
#SBATCH --time=1
 +
#SBATCH --job-name hello-world
 +
#SBATCH --nodes=5
 +
#SBATCH --mail-type=BEGIN,END,FAIL
 +
#SBATCH --mail-user=excellent_email_user@earlham.edu
 +
 +
echo "queue/partition is `echo $SLURM_JOB_PARTITION`"
 +
echo "running on `echo $SLURM_JOB_NODELIST`"
 +
echo "work directory is `echo $SLURM_SUBMIT_DIR`"
 +
 +
/bin/hostname
 +
srun -l /bin/hostname
 +
sleep 10
 +
srun -l /bin/pwd
 +
 +
</pre>
 +
 +
Interactive and CLI methods also exist.
 +
 +
== Conversion from Torque to Slurm ==
  
 
To submit a job to PBS, you'll need to write a shell script wrapper around it and submit it through qsub on your system of choice. For example (change the specific options):
 
To submit a job to PBS, you'll need to write a shell script wrapper around it and submit it through qsub on your system of choice. For example (change the specific options):
Line 106: Line 134:
 
</pre>
 
</pre>
  
−
= OLD: Using qsub =
+
= About qsub =  
−
 
 
−
Torque is our scheduler. It is gradually being phased out in favor of Slurm but remains installed for now. The command to submit something to be run on the scheduler is qsub. Knowing how to use qsub is the key to using our high-performance computing systems effectively.
 
−
 
 
−
== Submit job using a .qsub file (recommended) ==
 
−
 
 
−
To submit a job to PBS, you'll need to write a shell script wrapper around it and submit it through qsub on your system of choice. For example (change the specific options):
 
−
 
 
−
<pre>
 
−
#PBS -N out
 
−
#PBS -q batch
 
−
#PBS -l nodes=1:ppn=1
 
−
#PBS -m abe
 
−
#PBS -M username@example.com
 
−
#PBS -o /path/to/my/stdout
 
−
#PBS -e /path/to/my/stderr
 
−
</pre>
 
−
 
 
−
In this example:
 
−
* -N gives the job name
 
−
* -q tells Torque which queue to submit the job to
 
−
* -l tells Torque how many nodes and how many processors-per-node to request
 
−
** on a phat node, nodes=1; on a cluster, nodes=N where N>1 (most often)
 
−
* -m abe tells Torque what mail options to use. abe enables all mail information.
 
−
* -M tells Torque the email address
 
−
* -o tells Torque where to copy STDOUT to
 
−
* -e tells Torque where to copy STDERR to
 
−
 
 
−
It's sometimes useful to include the following snippet as well, so that your output files remind you of some other system details in the days or weeks after your job terminates:
 
−
 
 
−
<pre>
 
−
echo "running on `cat $PBS_NODEFILE`"
 
−
echo "hostname is `hostname`"
 
−
echo "on launch cwd is `pwd`"
 
−
echo "PBS_O_WORKDIR is `echo $PBS_O_WORKDIR`"
 
−
 
 
−
cd $PBS_O_WORKDIR
 
−
</pre>
 
−
 
 
−
Using a qsub file is the recommended approach because it is easy to gather your files and share them with others if you have an issue.
 
−
 
 
−
== Submit job via command line interface ==
 
−
 
 
−
You can also run a bash script through qsub via CLI. Here's an example script:
 
−
 
 
−
<tt>
 
−
#!/bin/sh
 
−
#PBS -N primes_block
 
−
#PBS -o /path/to/out.txt
 
−
#PBS -e /path/to/err.txt
 
−
#PBS -q batch
 
−
#PBS -m abe
 
−
 
−
echo "Hello world!" <br>
 
−
exit $! <br>
 
−
</tt>
 
−
 
 
−
* To submit that job, you might run something like this on b0:
 
−
<pre>$ qsub -l nodes=8,cput=0:30:0 ./hello.sh</pre>
 
−
This will request 8 nodes for a total of 30 minutes CPU time (granted, this is an absurdly long time to request for "hello world").
 
−
 
 
−
== Using qsub interactively ==
 
−
 
 
−
Often it is useful to test commands, scripts, etc. before submitting them to the queue. To do that you can launch qsub interactively and the scheduler will assign resources for use from a shell, e.g. this command would assign two nodes with two processors on each node: <pre>$ qsub -I -l nodes=2:ppn=2 </pre>
 
−
 
 
−
After launching you will be at a new shell prompt. The command: <pre>$ cat $PBS_NODEFILE</pre> will show you which node(s) were assigned to your interactive job.
 
−
 
 
−
== Related commands ==
 
−
 
 
−
Including qsub for search convenience, here are some common Torque commands you'll probably use (in descending order of probable frequency):
 
−
 
 
−
* qsub: Submit a job to a queue.
 
−
* qstat: Show a list of running, queued, or recently ended jobs.
 
−
* qdel: Delete a job you've submitted.
 
−
* qhold: Hold a job from execution.
 
  
−
Each command has a man page.
+
Before Slurm we used Torque and its associated software, including qsub. This is now deprecated and should not be used on the Earlham CS cluster systems.

Revision as of 14:05, 11 November 2020

This document presumes zero prior knowledge of cluster computing. If instead you're an intermediate user (e.g. you have an account and have run a few jobs before but need a reminder) the table of contents is your friend.

This document gives you all the information you need to choose a system, log in to a cluster/phat node, write a script, submit it via qsub to the scheduler, and find the output. As such, these notes cover hardware and software. (If you're a sysadmin, you may be interested in this page instead.)

Prerequisites

  1. Get a cluster account. You can email admin at cs dot earlham dot edu or a current CS faculty member to get started. Your user account will grant access to all the servers below, and you will have a home directory at ~username that you can access when you connect to any of them.
    1. Note: if you have a CS account, you will use the same username and password for your cluster account.
  2. Connect through a terminal via ssh to username@hopper.cluster.earlham.edu. If you intend to work with these machines a lot, you should also configure your ssh keys.

Cluster systems to choose from

The cluster dot earlham dot edu domain consists of clusters (a collection of physical servers linked through a switch to perform high-performance computing tasks with distributed memory) and jumbo servers (nee "phat nodes"; a system comprising one physical server with a high ratio of disk+RAM to CPU, good for jobs demanding shared memory).

Our current machines are:

  • whedon: newest cluster; 8 compute nodes; Torque-only pending an OS upgrade
  • layout: cluster; 4 compute nodes, pre-whedon, features NVIDIA GPGPU's and multiple CUDA options
  • lovelace: newest jumbo server
  • pollock: jumbo server, older than lovelace but well-tested and featuring the most available disk space

To get to, e.g., whedon, from hopper, run ssh whedon.

If you're still not sure, click here for more detailed notes.

Cluster software bundle

The cluster dot earlham dot edu servers all run a supported CentOS version.

All these servers (unless otherwise noted) also feature the following software:

  • Slurm (scheduler): submit a job with sbatch jobname.sbatch, delete it with scancel jobID. Running a job has its own doc section below.
  • Environment modules: run module avail to see available software modules and module load modulename to load one; you may load modules in bash scripts and qsub jobs as well.

The default shell on all these servers is bash.

The default Python version on all these servers is Python 2.x, but all have at least one Python 3 module with a collection of widely-used scientific computing libraries.

Using Slurm

Slurm is our batch scheduler.

You can check that it's working by running: srun -l hostname

You can submit a job in a script with the following: sbatch my_good_script.sbatch

Here's an example of a batch file: 

#!/bin/sh
#SBATCH --time=1
#SBATCH --job-name hello-world
#SBATCH --nodes=5
#SBATCH --mail-type=BEGIN,END,FAIL 
#SBATCH --mail-user=excellent_email_user@earlham.edu

echo "queue/partition is `echo $SLURM_JOB_PARTITION`"
echo "running on `echo $SLURM_JOB_NODELIST`"
echo "work directory is `echo $SLURM_SUBMIT_DIR`"

/bin/hostname
srun -l /bin/hostname
sleep 10
srun -l /bin/pwd

Interactive and CLI methods also exist.

Conversion from Torque to Slurm

To submit a job to PBS, you'll need to write a shell script wrapper around it and submit it through qsub on your system of choice. For example (change the specific options):


Commands
Torque Slurm Description
qsub sbatch run/submit a batch job
qstat squeue show jobs currently in the queue
qdel scancel cancel a job
pbsnodes -a scontrol show nodes show nodes in the cluster
Environment Variables
Torque Slurm Description
$PBS_QUEUE $SLURM_JOB_PARTITION the queue/partition you are in
cat $PBS_NODEFILE $SLURM_JOB_NODELIST there's no equivalent of the nodes file but there is an environment variable that stores that information
$PBS_O_WORKDIR $SLURM_SUBMIT_DIR working directory from which the command was run

Example script 

#!/usr/bin/bash

#SBATCH --job-name hello-world
#SBATCH --nodes=5
#SBATCH --mail-type=BEGIN,END,FAIL 
#SBATCH --mail-user=excellent_email_user@earlham.edu

echo "queue is `echo $SLURM_JOB_PARTITION`"
echo "running on `echo $SLURM_JOB_NODELIST`"
echo "work directory is `echo $SLURM_SUBMIT_DIR`"

srun -l echo "hello world!"

About qsub

Before Slurm we used Torque and its associated software, including qsub. This is now deprecated and should not be used on the Earlham CS cluster systems.

Retrieved from "https://wiki.cs.earlham.edu/index.php?title=Getting_started_on_clusters&oldid=17722"