Is your feature request related to a problem? Please describe.
Slurm support multiple controllers for HA.
Add support for multiple controllers with each in separate AZs.
Describe the solution you'd like
Currently the slurm config, binaries, and state save location is stored on the EBS volumes of the controller.
This would need to be move to a multi-AZ file system such as EFS or FSxN.
EFS may not have the IOPS required by to keep the state save location from becoming a bottleneck.
Add a configuration parameter for the head node to allow 1-3 controllers.
Create the specified number of controllers and update slurm.conf appropriately.
Is your feature request related to a problem? Please describe. Slurm support multiple controllers for HA. Add support for multiple controllers with each in separate AZs.
Describe the solution you'd like Currently the slurm config, binaries, and state save location is stored on the EBS volumes of the controller. This would need to be move to a multi-AZ file system such as EFS or FSxN. EFS may not have the IOPS required by to keep the state save location from becoming a bottleneck. Add a configuration parameter for the head node to allow 1-3 controllers. Create the specified number of controllers and update slurm.conf appropriately.