Jobs won't run on more than one node

Audience: Faculty, Postdocs, Researchers, Staff and Students

This KB Article References: High Performance Computing
This Information is Intended for: Faculty, Postdocs, Researchers, Staff, Students
Last Updated: June 28, 2018

PBS Torque, our job scheduler, uses SSH to distribute jobs to the compute nodes. In order for keyless SSH access to work a few things need to be in place:

  • Permissions on the user’s home directory (e.g. /gpfs/home/jsmith) must ONLY give the owner (jsmith) write permissions (e.g. drwxr-x---)
  • Permissions on the user’s ssh directory (e.g. /gpfs/home/jsmith/.ssh) must ONLY give the owner read & write permissions (e.g. drwx------)
  • Permissions on the authorized_keys file (e.g. /gpfs/home/jsmith/.ssh/authorized_keys) must only give the owner read & write permissions (e.g. -rw-------)
  • The user's authorized_keys file must contain ~/.ssh/id_ecdsa.pub on its own line
  • The ssh key for accessing resources within the cluster (e.g. /gpfs/home/jsmith/.ssh/id_rsa) must NOT have a passphrase

The above are all part of the default configuration on our clusters. Users are advised not to change these settings.

SUBMIT A TICKET

Additional Information


There are no additional resources available for this article.

Getting Help


The Division of Information Technology provides support on all of our services. If you require assistance please submit a support ticket through the IT Service Management system.

Submit A Ticket

For More Information Contact


IACS Support System