Cluster scaling-down refers to the process of reducing Task nodes through the console when there is an excess of computing resources in the EMR cluster. Moreover, when the cluster has routing nodes (router nodes) that do not share the load of the Master node, or serve as the task submission machine for the cluster, the cluster scaling-down function can be achieved by reducing nodes.
Preparations
Core nodes store data, and it is not recommended to terminate them after expansion to avoid potential data security risks. Once a node is terminated, the data on that node will not be retained. If you choose to scale down (terminate) nodes, it implies that you have confirmed that the data on the selected nodes can be destroyed.
For annual and monthly subscription clusters: After termination, the cluster will be retained in the recycle bin for 7 days, during which you can choose to restore the cluster. After 7 days, the cluster will be permanently destroyed and cannot be restored. Please proceed with caution.
For pay-as-you-go clusters: After termination, the cluster will not be retained in the recycle bin and will be permanently destroyed, unable to be restored. Please proceed with caution.
Note
Ensure that you have backed up your data before terminating the cluster, as data cannot be retrieved once the cluster is destroyed.
Instructions
1. Log in to the EMR Console, and in the cluster list, click on the corresponding cluster ID/Name to access the cluster details page.
2. Within the cluster details page, select Cluster Resources > Resource Management. In the node list, select the cloud resources that support scaling down (task nodes or router nodes). When other types of nodes are selected concurrently, the Scale Down option cannot be clicked.
3. In the pop-up window, you need to confirm the termination (scaling down) of instances. If the cluster's meteDB is a shared meta-database, it will be retained and you will need to manually operate it in the CDB console.
