Scheduling a Job using the SLURM job scheduler

Intro to SLURM slides

The Computer Science Department uses a “job scheduler” called SLURM. The purpose of a job scheduler such as SLURM is to allocate computational resources to users who submit “jobs” to a queue. The job scheduler looks at the requirements stated in the job script and allocates a server (or servers) to the job which matches the requirements specified in the job script.

Due to the way sbatch spawns a bash session (non-login session), some init files are not loaded from /etc/profile.d. This prevents the initialization of the Environment Modules system and will prevent you from loading software modules.

To fix this, simply include the following line in your sbatch scripts:

source /etc/profile.d/modules.sh

Slurm Commands Cheat Sheet

SLURM provides a set of tools to use for interacting with the scheduling system. To view information about compute nodes in the SLURM system, we can use the command sinfo.

pgh5a@portal01 ~ $ sinfo
PARTITION     AVAIL  TIMELIMIT  NODES  STATE NODELIST
main*            up   infinite     37   idle hermes[1-4],artemis[1-7],slurm[1-5],nibbler[1-4],trillian[1-3],granger[1-6],granger[7-8],ai0[1-6]
qdata            up   infinite      8   idle qdata[1-8]
qdata-preempt    up   infinite      8   idle qdata[1-8]
falcon           up   infinite     10   idle falcon[1-10]
intel            up   infinite     24   idle artemis7,slurm[1-5],granger[1-6],granger[7-8],nibbler[1-4],ai0[1-6]
amd              up   infinite     13   idle hermes[1-4],artemis[1-6],trillian[1-3]

With sinfo we can see a listing of what SLURM calls partitions and a list of nodes associated with these partitions. A partition is a grouping of nodes, for example our main partition is a group of all SLURM nodes that are not reserved and can be used by anyone. Notice that hosts can be listed in several different partitions.. For example, slurm[1-5] can be found in both the main and intel partitions. Intel is a partition of systems with Intel processors, likewise the hosts in amd have AMD processors.

To view jobs running on the queue, we can use the command squeue. Say we have submitted one job to the main partition, running squeue will look like this:

pgh5a@portal01 ~ $ squeue
             JOBID PARTITION     NAME     USER ST       TIME  NODES NODELIST(REASON)
            467039      main    sleep    pgh5a  R       0:06      1 artemis1

and now that a node has been allocated, that node artemis1 will show as alloc in sinfo

pgh5a@portal01 ~ $ sinfo
PARTITION     AVAIL  TIMELIMIT  NODES  STATE NODELIST
main*            up   infinite     37   idle hermes[1-4],artemis[2-7],slurm[1-5],nibbler[1-4],trillian[1-3],granger[1-6],granger[7-8],ai0[1-6]
main*            up   infinite      1  alloc artemis1
qdata            up   infinite      8   idle qdata[1-8]
qdata-preempt    up   infinite      8   idle qdata[1-8]
falcon           up   infinite     10   idle falcon[1-10]
intel            up   infinite     24   idle artemis7,slurm[1-5],granger[1-6],granger[7-8],nibbler[1-4],ai0[1-6]
amd              up   infinite     13   idle hermes[1-4],artemis[1-6],trillian[1-3]

To use SLURM resources, you must submit your jobs (program/script/etc.) to the SLURM controller. The controller will then send your job to compute nodes for execution, after which time your results will be returned.

Users can submit SLURM jobs from any of the power servers: power1-power6. From a shell, you can submit jobs using the commands srun or sbatch. Let's look at a very simple example script and sbatch command.

Here is our script, all it does is print the hostname of the server running the script. We must add SBATCH options to our script to handle various SLURM options.

#!/bin/bash
 
#SBATCH --job-name="Slurm Simple Test Job" #Name of the job which appears in squeue
#
#SBATCH --mail-type=ALL
#SBATCH --mail-user=pgh5a@virginia.edu
#
#SBATCH --error="my_job.err"                    # Where to write std err
#SBATCH --output="my_job.output"                # Where to write stdout
#SBATCH --nodelist=slurm1
 
hostname

Let's put this in a directory called slurm-test in our home directory. We run the script with sbatch and the results will be put in the file we specified with --output. If no output file is specified, output will be saved to a file with the same name as the SLURM jobid.

pgh5a@power3 ~ $ cd slurm-test/ 
pgh5a@power3 ~/slurm-test $ chmod +x test.sh 
pgh5a@power3 ~/slurm-test $ sbatch test.sh 
Submitted batch job 466977
pgh5a@power3 ~/slurm-test $ ls
my_job.err  my_job.output  test.sh
pgh5a@power3 ~/slurm-test $ cat my_job.output 
slurm1

Here is a similar example using srun running on multiple nodes:

pgh5a@power3 ~ $ srun -w slurm[1-5] -N5 hostname
slurm4
slurm1
slurm2
slurm3
slurm5

Please be aware of jobs you start and make sure that they finish executing. If your job does not converge, it will sit in the queue taking up resources and preventing others from running their jobs.

To cancel a running job, use the scancel [jobid] command

ktm5j@power1 ~ $ squeue
             JOBID PARTITION     NAME     USER ST       TIME  NODES NODELIST(REASON)
            467039      main    sleep    ktm5j  R       0:06      1 artemis1           <--  Running job
ktm5j@power1 ~ $ scancel 467039

Slurm refers to job queues as partitions. These queues can have unique constraints such as compute nodes, max runtime, resource limits, etc. There is a main queue, which will make use of all non-GPU compute nodes.

Partition is indicated by -p partname or --partition partname.

To specify a partition with sbatch file:

#SBATCH --partition=gpu

Or from the command line with srun

-p gpu

An example running the command hostname on the intel partition, this will run on any node in the partition:

srun -p intel hostname

Slurm handles GPUs and other non-CPU computing resources using what are called GRES Resources (Generic Resource). To use the GPU(s) on a system using Slurm, either using sbatch or srun, you must request the GPUs using the --gres:x option. You must specify the gres flag followed by : and the quantity of resources

Say we want to use 4 GPUs on a system, we would use the following sbatch option:

#SBATCH --gres=gpu:4

Or from the command line

--gres=gpu:4

We can use srun to spawn an interactive shell on a SLURM compute node. While this can be useful for debugging purposes, this is not how you should typically use the SLURM system. To spawn a shell we must pass the --pty option to srun so output is directed to a pseudo-terminal:

ktm5j@power3 ~/slurm-test $ srun -w slurm1 --pty bash -i -l -
ktm5j@slurm1 ~/slurm-test $ hostname
slurm1
ktm5j@slurm1 ~/slurm-test $

The -i argument tells bash to run as interactive. The -l arg instructs bash that this is a login shell, this, along with the final - are important to reset environment variables that otherwise might cause issues using Environment Modules

If a node is in a partition other than the default “main” partition (for example, the “gpu” partition), then you must specify the partition in your command, for example:

ktm5j@power3 ~/slurm-test $ srun -w lynx05 -p gpu --pty bash -i -l -

Reservations for specific resources or nodes can be made by submitting a request to cshelpdesk@virginia.edu. For more information about using reservations, see the main article on SLURM Reservations