With t2.micro the process of configuring Cloudera Manager will not be able to succeed. That instance type simply doesn't have enough memory for the main server process and all the management services. Our recommendation is to use m4.large or m4.xlarge. Also for Director itself you should use an instance like c3.large for best performance.    Regarding the operation system our recommendation is to use the official releases either as community AMIs or from the AWS Marketplace: https://aws.amazon.com/marketplace/seller-profile?id=16cb8b03-256e-4dde-8f34-1b0f377efe89 (for CentOS)   This documentation page contains some more instructions on how to find an AMI: http://www.cloudera.com/content/www/en-us/documentation/director/latest/topics/director_deployment_ami.html   Also see Requirements and Supported Versions for additional information: http://www.cloudera.com/content/www/en-us/documentation/director/latest/topics/director_deployment_requirements.html  ... View more

m3.medium is definitely too small to run Cloudera Manager and all the management services.    My recommendation is to try again with m4.large or m4.xlarge as the Cloudera Manager instance. That should take memory related failures out of the picture.  ... View more

There is no such AMI matained by Cloudera. You can install R using a custom bootstrap script that Director runs or build your own AMI with those bits pre-installed.  ... View more

One suggestion is to give it  a try with gp2 for the root disk volume. That may make a difference.    There is no explicit way of increasing that timeout. It can be done by increasing the agent heartbeat but that has important implications on other core Cloudera Manager services and features. That timeout during FirstRun is usually a sign of poor AWS performance. We faced a similar challenge in the past in our own testing and had to find ways to use different AWS accounts to get better isolation.  ... View more

Hi Kirk -    > 1.  The application log doesn't reveal the aws CLI commands that are run.  Is there a way to  > see those commands?  Either before (as a dry run feature) or after?    Cloudera Director integrates with the AWS API using the Java SDK. There is no straightforward way to translate those method calls as AWS CLI commands you can run standalone. In theory you should be able to figure out how to use the AWS CLI to start similar instances by inspecting the existing ones.    > 2. The nodes that get provisioned do not have anything in the /etc/hosts file.  How are they resolving each other?   > Typically, the hosts file configuration is critical to getting Cloudera Manager to run successfully.   We expect DNS configuration to be handled outside of Cloudera Director using different mechanisms: VPC configuration to use AmazonProvidedDNS via DHCP Options Sets custom bootstrap script that Director runs to configure the hostname scripts embedded into a custom AMI that configure the proper hostname during boot and handle DNS registration Cloudera Director doesn't attempt to perform any DNS configs.  ... View more