Among the many recent updates to CISL documentation as the PBS Pro workload manager replaced Slurm on the Casper cluster, staff members suggested highlighting two for special mention:
- Users who are logged in on Cheyenne or Casper can now query jobs running on the other system by appending @cheyenne or @casper to a qstat command to specify where the job is running. Get the details here.
- The experts recommend not using the -V option to propagate all environment settings to a batch job. Why? It can cause odd behaviors and job failures if you’re submitting a Casper job from Cheyenne. See this page for details.