Last modified August 16, 2021

gsctl scale cluster

gsctl and the REST API are being phased out. We don't have an end-of-life date yet. However, we recommend to familiarize yourself with our Management API and the kubectl gs plugin as a future-proof replacement.

The gsctl scale cluster command allows to specify the number of worker nodes for an on-premises (KVM) cluster.

For clusters on AWS and Azure, instead of scaling an entire cluster, please see gsctl update nodepool regarding how to scale a node pool.

Notes on worker node removal

When reducing the worker node count, you have no influence in which exact order worker nodes are removed. Your workloads have to be configured in a way that single pods can be removed any time. See our article on recommendations and best practices for details on how to achieve that.

Command usage

To scale a cluster to a specific number of nodes, e. g. 5, use this syntax:

gsctl scale cluster f0r14 --num-workers 5

You can also use the cluster’s name for identifying the cluster:

gsctl scale cluster "Cluster name" --num-workers 5

After the command has been executed, it can take a few minutes until the worker node count has adapted to the new settings.

Confirmation

When reducing the number of worker nodes, you will have to confirm the command execution in an interactive prompt. This prompt can be suppressed using the --force flag.

When adding worker nodes, no such confirmation is required.

Full argument reference

  • -w, --num-workers: The intended number of worker nodes.
  • --force: If set, no confirmation is required when reducing the number of workers. You should only use this argument in automations when you are sure that reducing the number of workers is desired.

Use gsctl scale cluster --help for a additional (global) arguments.