Node Cordoning and Uncordoning

Node cordoning is the process of marking compute nodes as unschedulable, preventing new workflows from being placed on them. This is commonly used during maintenance, node upgrades, or when preparing to remove nodes from the cluster.

Prerequisites

Changes to compute node scheduling require that a fuzzball admin context is set up and authorized.

$ fuzzball admin context create <context_name> <api_url>

$ fuzzball admin context login -u <admin_username> -p '<admin_password>'

Cordon Node

The fuzzball admin scheduler node cordon command marks one or more compute nodes as unschedulable.

Cordon a Single Node

$ fuzzball admin scheduler node cordon <node-id>

Example:

$ fuzzball admin scheduler node cordon node-worker-01

Cordon All Nodes

$ fuzzball admin scheduler node cordon --all

Alternative syntax:

$ fuzzball admin scheduler node cordon -a

Uncordon Node

The fuzzball admin scheduler node uncordon command marks one or more compute nodes as schedulable again.

Uncordon a Single Node

$ fuzzball admin scheduler node uncordon <node-id>

Example:

$ fuzzball admin scheduler node uncordon node-worker-01

Uncordon All Nodes

$ fuzzball admin scheduler node uncordon --all

Alternative syntax:

$ fuzzball admin scheduler node uncordon -a