We are using SLURM (Simple Linux Utility for Resource Management) for our resource manager and job scheduler. Below are several of the basic commands you will need to interact with the cluster.
In a nutshell, sbatch and salloc allocate resources to the job, while srun launches parallel tasks across those resources. When invoked within a job allocation, srun will launch parallel tasks across some or all of the allocated resources. In that case, srun inherits by default the pertinent options of the sbatch or salloc which it runs under. You can then (usually) provide srun different options which will override what it receives by default. Each invocation of srun within a job is known as a job step.
srun can also be invoked outside of a job allocation. In that case, srun requests resources, and when those resources are granted, launches tasks across those resources as a single job and job step.
Source: slurm-devel mailing list post
Information on jobs
List all current jobs for a user:
squeue -u <username>
List all running jobs for a user:
squeue -u <username> -t RUNNING
List all pending jobs for a user:
squeue -u <username> -t PENDING
List detailed information for a job (useful for troubleshooting):
scontrol show jobid -dd <jobid>
To cancel one job:
To cancel all the jobs for a user:
scancel -u <username>
To cancel all the pending jobs for a user:
scancel -t PENDING -u <username>
To cancel one or more jobs by name:
scancel --name myJobName
To pause a particular job:
scontrol hold <jobid>
To resume a particular job:
scontrol resume <jobid>
To requeue (cancel and rerun) a particular job:
scontrol requeue <jobid>
sbatch submits a batch script to SLURM. The batch script may be givento sbatch through a file name on the command line, or if no file name is specified, sbatch will read in a script from standard input. The batch script may contain options preceded with #SBATCH before any executable commands in the script.
sbatch exits immediately after the script is successfully transferred to the SLURM controller and assigned a SLURM job ID. The batch script is not necessarily granted resources immediately, it may sit in the queue of pending jobs for some time before its required resources become available.
When the job allocation is finally granted for the batch script, SLURM runs a single copy of the batch script on the first node in the set of allocated nodes.
$ cat sample.txt #!/bin/sh #SBATCH --mail-type=ALL #SBATCH --email@example.com #SBATCH --ntasks=5 #SBATCH --job-name=test srun R CMD BATCH script.R $ sbatch ./sample.txt Submitted batch job 25618 $ cat slurm-25618.out # any output to STDOUT would be in this file
Array jobs provide a mechanism for submitting and managing collections of jobs. Job arrays are only supported from batch jobs. To control the size of the array the –array or -a option is passed to sbatch.
An example batch script would look like
$ cat sample.txt #!/bin/sh #SBATCH --mail-type=ALL #SBATCH --firstname.lastname@example.org #SBATCH --time=1-0 #SBATCH --array=1-100 srun R CMD BATCH ./script.R
This command is used to view information about the slurm scheduling queue.
To view you jobs in the queue regardless of their current state.
$ squeue -u $USER
Job Status Codes
Typically your job will be either in the Running state of PenDing state. However here is a breakdown of all the states that your job could be in.
|CA||CANCELLED||Job was explicitly cancelled by the user or system administrator. The job may or may not have been initiated.|
|CD||COMPLETED||Job has terminated all processes on all nodes.|
|CF||CONFIGURING||Job has been allocated resources, but are waiting for them to become ready for use (e.g. booting).|
|CG||COMPLETING||Job is in the process of completing. Some processes on some nodes may still be active.|
|F||FAILED||Job terminated with non-zero exit code or other failure condition.|
|NF||NODE_FAIL||Job terminated due to failure of one or more allocated nodes.|
|PD||PENDING||Job is awaiting resource allocation.|
|R||RUNNING||Job currently has an allocation.|
|S||SUSPENDED||Job has an allocation, but execution has been suspended.|
|TO||TIMEOUT||Job terminated upon reaching its time limit.|
Used to signal jobs or job steps that are under the control of Slurm.
scancel is used to signal or cancel jobs or job steps. An arbitrary number of jobs or job steps may be signaled using job specification filters or a space separated list of specific job and/or job step IDs. A job or job step can only be signaled by the owner of that job or user root. If an attempt is made by an unauthorized user to signal a job or job step, an error message will be printed and the job will not be signaled.
$ squeue JOBID PARTITION NAME USER ST TIME NODES NODELIST(REASON) 29908 biostat-d bash schelcj R 0:05 2 cn[001-002] $ scancel 29908 $ squeue JOBID PARTITION NAME USER ST TIME NODES NODELIST(REASON) $
Here we see our jobid is 29908 then we cancel that job with the scancel command.
Used to launch a jobs in the cluster
The following example runs the /bin/hostname command on at least two nodes in the cluster. STDOUT/STDERR return to your terminal.
$ srun --nodes=2 hostname cn001 cn002
You can specify the number of cpu’s to use per job with the –cpus-per-task option.
$ srun --nodes=2 --cpus-per-task=3 R CMD BATCH script.R
All SAS jobs should be run in the biostat-sas partition to meet license restrictions.
$ srun --partition=biostat-sas some_sas_script.sas
Provided there are available licenses the job will be launched according the the normal allocation process. If the license pool has been exhausted then the job will remain in the pending state until a license is freed for use.
Obtain an interactive SLURM job allocation (a set of nodes), execute a command, and then release the allocation when the command is finished.
This just requests an allocation of cluster resources and gives you an interactive shell (when you exit that shell the allocation is relinquished and all jobs running are cancelled.) For example
$ salloc -N2 salloc: Granted job allocation 29903 $ srun hostname cn002 cn001 $
Here we have requested two nodes be allocated then we run /bin/hostname to see what nodes we were allocated. Any command we run within this allocation is run in parallel across the number of nodes we requested, in this case 2. So for instance if you had an R script that generated its own unique dataset for each invocation you could have ran that script across two nodes. For example;
$ salloc -N2 salloc: Grant job allocation 12345 $ srun R CMD BATCH script.R & $ srun R CMD BATCH script.R & $ srun R CMD BATCH script.R & $
In this example you have a two node allocation with each call to srun using a single core on both machines. Each call to srun launches the same R script on each node and the & sign backgrounds the process so that you can issue more srun commands. This example is now running the R script 6 times across two nodes.
$ env|grep SLURM SLURM_NODELIST=cn[001-002] SLURM_NNODES=2 SLURM_JOBID=29903 SLURM_TASKS_PER_NODE=2(x2) SLURM_JOB_ID=29903 SLURM_JOB_NODELIST=cn[001-002] SLURM_JOB_CPUS_PER_NODE=2(x2) SLURM_JOB_NUM_NODES=2
While within the interactive allocation you have several environmental varilables available to you that describe the allocation. Having things like the JOB_ID available make it possible to use the –dependency option to srun.
For times when you need an interactive shell to debug code, do post processing, run tests, or do anything that would be outside the appropriate use of the login nodes you can submit an interactive job that will give you a shell on a compute node. To submit an interactive job to slurm use the salloc command and the srun command in combination. Here is an example:
$ salloc --time=1:00:00 srun --pty /bin/bash salloc: Pending job allocation 12000 salloc: job 12000 queued and waiting for resources salloc: job 12000 has been allocated resources salloc: Granted job allocation 12000 schelcj@cn004:~$
This will give you a shell for one hour. Normal time argruments apply. From here you could for instance start an R session to debug some of your code.