You might need to set your node as a manager node in the following situations:
- Cluster maintenance: You might need to take some manager nodes offline for maintenance. Before the maintenance, you can set a worker node to a manager node to make sure the normal operation of the cluster without interrupting business.
- Resource allocation optimization: You might want to set an Elastic Compute Service (ECS) instance with better resource configuration as a manager node to optimize the resource allocation.
- Build a cluster architecture with higher availability: You might need to upgrade the existing cluster to improve the error tolerance of the cluster. To achieve this, set a worker node to a manager node to build a high-availability cluster with more manager nodes (2n+1 manager nodes, n≥1). For example, to expand a cluster with three manager nodes to one with five manager nodes, you must manually set the newly added two nodes to manager nodes.
Note: It does not mean that the more manager nodes, the better. The Docker official recommendation is a maximum of seven manager nodes.
By default, all nodes are
ACTIVE and manager nodes are also assigned with tasks. However, you might want to assign only the management and scheduling tasks to manager nodes out of cluster business management and maintenance considerations. To achieve this, you can set the node to
Drain to make it exclusively manage the containers. For more information, see Drain a node on the swarm.
Log on to the Container Service console.
Click Nodes under Swarm in the left-side navigation pane.
Select the cluster in which the node resides from the Cluster list.
Click More at the right of the node and then click Promote from the list.
In the confirmation dialog box, click Confirm.