In practical application, there may be instances where nodes within the cluster, particularly the MASTER nodes, are deficient in CPU or memory. In such cases, the CPU or memory size can be increased by modifying the configuration. This document provides an overview of the relevant operations for changing instance configurations through the EMR Console.
Note
During the process of node configuration changes, the system will shut down. This shutdown may potentially disrupt the normal operation of the cluster, and could even result in business interruption. Please proceed with caution after careful evaluation.
During node reconfiguration, changes to the size of the data disk and system disk are not supported.
Preparations
1. After reconfiguration of pay-as-you-go nodes, the billing tier will reset to the first tier. For annual and monthly subscription clusters, the corresponding fee difference needs to be compensated.
2. Local disk models, POD resources, and spot instance billing models do not support reconfiguration.
3. Costs in a batch configuration change will be deducted item by item. Please ensure your account balance is sufficient.
4. The refund amount will be returned to your Tencent Cloud account in accordance with the ratio of cash and gift money used at the time of purchase; discounts and vouchers used during purchase are non-refundable.
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. In the cluster details page, select Cluster Resources to enter the resource management page. Based on business needs, select the nodes that need to be changed and perform the Change Configuration Operation; batch changes are supported, but only nodes of the same billing type can be changed to a uniform configuration during a batch change.
3. On the configuration adjustment page, confirm the relevant change information. It is crucial to read the important prompts carefully and check the box to agree to the change information.
4. On the target configuration page, select configurations such as machine type, instance type, and machine type list. After confirming the cost is correct, click Confirm Changes to adjust the configuration.
5. The cost details page will separately display the fees required to change different nodes to a uniform configuration.
6. (Optional) After changing the configuration, if you need to adjust the resources of the corresponding components, you will need to reissue the configuration in configuration management and restart the service.
Note
Yarn resources are automatically adjusted according to machine specifications by default. After reconfiguration, the resource size changes with the reconfiguration specifications, eliminating the need for manual configuration adjustments.
If you have manually adjusted the Yarn resource configuration, after reconfiguration, you need to go to Configuration Management to modify the parameter values of yarn.nodemanager.resource.cpu-vcores and yarn.nodemanager.resource.memory-mb. Click Save Configuration, and after issuing the configuration, restart the NodeManager service to complete the Yarn resource configuration update.
