Your First Job

Step-by-step walkthrough for submitting your first job on either cluster.

This is how you test your code on the cluster! You don’t run scripts directly on the login node - you submit them as jobs. Even small tests should be submitted to the scheduler. This teaches you the submission process you’ll use for all your computational work.

Choose Your Cluster

Select the guide for your assigned cluster:

Your First Job on OrangeGrid

Step 1: Connect to OrangeGrid 

ssh netid@its-og-login3.syr.edu

(Use your assigned login node from your welcome email)

Step 2: Create Your Script 

nano hello.sh

Add this content:

#!/bin/bash
set -e

echo "Hello from OrangeGrid!"
echo "Host name: $(hostname)"
echo "Date: $(date)"
echo "Kernel: $(uname -r)"

# Do some work
echo "Calculating..."
python3 -c "print('Sum:', sum(range(1000000)))"

echo "Job complete!"
exit 0

Make it executable:

chmod +x hello.sh

Step 3: Create HTCondor Submit File 

nano hello.sub

Add this content:

# Point to your executable script
executable = hello.sh

# Output files
output = hello.$(cluster).$(process).out
error = hello.$(cluster).$(process).err
log = hello.$(cluster).log

# Submit one job
queue 1

Step 4: Submit the Job 

condor_submit hello.sub

You’ll see:

Submitting job(s).
1 job(s) submitted to cluster 54321.

Step 5: Check Job Status 

# Check your jobs
condor_q

# Check specific user
condor_q netid

# Watch continuously (update every 5 seconds)
watch -n 5 condor_q netid

Understanding Job States:

  • I - Idle (waiting in queue)
  • R - Running
  • H - Held (problem requiring manual intervention)
  • C - Completed

Step 6: View Results 

# View output
cat hello.54321.0.out

# Check for errors
cat hello.54321.0.err

Common HTCondor Commands:

  • condor_q - View job queue
  • condor_rm 54321 - Remove job 54321
  • condor_status - View available machines
  • condor_userprio - See who’s using the pool

Your First Job on Zest

Step 1: Connect to Zest 

ssh netid@its-zest-login1.syr.edu

(Use your assigned login node from your welcome email)

Step 2: Create Your Python Script 

nano hello_cluster.py

Add this content:

#!/usr/bin/env python3
import socket
import datetime

print("Hello from the cluster!")
print(f"Running on: {socket.gethostname()}")
print(f"Time: {datetime.datetime.now()}")

# Do some computation
result = sum(range(1000000))
print(f"Computation result: {result}")

# Save to file
with open('hello_output.txt', 'w') as f:
    f.write(f"Job ran on {socket.gethostname()}\n")
    f.write(f"Result: {result}\n")

Save with Ctrl+O, exit with Ctrl+X.

Step 3: Create Job Submission Script 

nano submit_hello.sh

Add this content (replace netid with your actual NetID):

#!/bin/bash
#SBATCH --nodes=1
#SBATCH --ntasks=1
#SBATCH --cpus-per-task=1
#SBATCH --time=00:05:00
#SBATCH --mail-type=ALL
#SBATCH --mail-user=netid@syr.edu

# Load Python environment
module load anaconda3

# Run the script
python hello_cluster.py

Step 4: Submit the Job 

sbatch submit_hello.sh

You’ll see output like:

Submitted batch job 12345

Step 5: Check Job Status 

# Check your jobs
squeue -u netid

# See detailed info
scontrol show job 12345

Step 6: View Results

Once the job completes, check your output:

# View the Slurm output file
cat slurm-12345.out

# View your custom output
cat hello_output.txt

Common Slurm Commands:

  • squeue - View job queue
  • scancel 12345 - Cancel job 12345
  • sinfo - View node information
  • sacct - View job accounting info

Next Steps

Now that you’ve submitted your first job, you can:

  1. Explore examples - Check the OrangeGrid or Zest repositories
  2. Set up your environment - See the Software & Environments guide
  3. Adapt examples - Modify example scripts for your research
  4. Optimize - Learn from resource usage and adjust requests

Need help? Email researchcomputing@syr.edu