Next Thurs. 2017-11-16 is a regularly scheduled maintenance window. We'd like to extend this from 2 to 45 hrs, mostly effecting PDAC. That would be 07:00-11:00 ProjTime. Any objections?
As part of NCSA's production framework development activities, we'll begin implementation of multiple SLURM queues on the verification cluster
Currently using 2 nodes for experimentation. This Fri., we anticipate increasing queue resources to 14 nodes, at least through the end of Nov. ( there are 48 nodes, each with 48 CPUs)
Initially, we'll implement & test the following queue structure:
Development queue: 34 nodes
Production queue: 14 nodes reserved with some ability to utilize (and release) idle development nodes
dev environment
Requests for change
Configure SLURM to accept jobs to use only partial nodes
DM developers need a build/test environment that supports docker containers
