Scaling an AWS cluster
The DC/OS AWS CloudFormation template is optimized to run DC/OS, but you might want to change the number of agent nodes based on your needs.
Important: Scaling down your AWS cluster could result in data loss. It is recommended that you scale down by 1 node at a time, letting the DC/OS service recover. For example, if you are running a DC/OS service and you scale down from 10 to 5 nodes, this could result in losing all the instances of your service.
To change the number of agent nodes with AWS:
- From AWS CloudFormation Management page, select your DC/OS cluster and click Update Stack.
- Click through to the Specify Parameters page, and you can specify new values for the PublicSlaveInstanceCount and SlaveInstanceCount.
- On the Options page, accept the defaults and click Next. Tip: You can choose whether to rollback on failure. By default this option is set to Yes.
- On the Review page, check the acknowledgement box and then click Create.
Your new machines will take a few minutes to initialize; you can watch them in the EC2 console. The DC/OS web interface will update as soon as the new nodes register.
Upgrading a DC/OS cluster in AWS
You can update an existing DC/OS cluster or services to use the latest DC/OS template.
To upgrade a DC/OS cluster:
Create a new DC/OS cluster by using the latest DC/OS template for AWS.
Migrate your active DC/OS services and apps to the new DC/OS cluster:
Migrate, Extract, Transform and Load (ETL) the app data to the new cluster.
Migrate your DC/OS services and apps to the new cluster.
Change the DNS so that it points to the DC/OS services running in the new cluster.
Shutdown your existing DC/OS cluster.