Multiple subnets in ParallelCluster

Question

Is there anyway to deploy compute accross Availibility Zones of a region?
It seems we can only specify one subnet per queue.
Also if I try to setup queues in different subnets I get 'The SubnetIds used for all of the queues should be the same.'.

I understand latency can get affected, but if the workload is not using networks, this would be advantageous to get access to more compute!

Accepted Answer

We're tracking this feature request in this GitHub issue: https://github.com/aws/aws-parallelcluster/issues/3576. We don't yet have a timeline to share for this feature enhancement.

If you need different "compute" you can define multiple [ComputeResources](https://docs.aws.amazon.com/parallelcluster/latest/ug/Scheduling-v3.html#Scheduling-v3-SlurmQueues-ComputeResources) or multiple [SlurmQueues](https://docs.aws.amazon.com/parallelcluster/latest/ug/Scheduling-v3.html#Scheduling-v3-SlurmQueues), even if they are in the same subnet.

An high level alternative is to use Slurm Federation to manage multiple Slurm clusters. See Workshop: [Using AWS ParallelCluster for Research](https://catalog.us-east-1.prod.workshops.aws/workshops/d431e0b1-9f08-4d82-822e-ea56962b2a0b/en-US/summary).

Enrico

Multiple subnets in ParallelCluster

Relevanter Inhalt