Troubleshooting Slurm reboot issues in AWS PCS
When you encounter node reboot problems, first check the node status using scontrol
show node . Then examine CloudWatch logs for both Slurm
(slurmctld and slurmd) and system logs to identify potential errors.nodename
For basic troubleshooting, verify network connectivity, check security group settings, and ensure all required services are running after the reboot. If problems persist after basic troubleshooting steps, contact AWS Support. When reaching out to support, provide relevant log excerpts, node status information, and a timeline of the reboot attempt to help expedite the resolution process.
Additional resources
-
For monitoring AWS PCS instances using CloudWatch, see Monitoring AWS PCS instances using Amazon CloudWatch.
-
For general troubleshooting, see Troubleshooting problems in AWS Parallel Computing Service.
-
For Slurm documentation, see Slurm Troubleshooting Guide
.
