Virtual Warehouse Impala - CDP Public Cloud (AWS)

[#1]

Infrastructure

Virtual Warehouse Impala will be running under Kubernetes Pods. The user can find the Kubernetes cluster running in the AWS account. The pods will be running in one of the K8s nodes.

The JDBC URL can be found from the Virtual Warehouse. 

Example:

jdbc:impala://<IMPALA_Coordinator_URL>:443/default;AuthMech=3;transportMode=http;httpPath=cliservice;ssl=1;UID=xxxx;PWD=PASSWORD

The server name would be <IMPALA_Coordinator_URL> and the port is 443.

Access to Cloudera Manager, other cluster UIs, and endpoints (such as JDBC) are always through the secure Knox gateway on port 443. Users are automatically logged in with their CDP credentials.

If you get "Error while trying to connect" message from the visualization tool, check if port 443 is allowed for Impala coordinator and Knox gateway server from the server where the visualization tool is hosted.

Have come across some visualization tools like DOMO; its connectors can only talk to public IPs, thus we would recommend the user to use either of the networking infras to configure (CDW):

1. Public Load Balancer, Public Worker Nodes > All AWS network components will have a publicly visible IP address assigned to them.
2. Public Load Balancer, Private Worker Nodes > All traffic is routed from the load balancer (Amazon ELB), which is located in a public subnet and thus will have a public-facing IP address.

If users go for option 3. (Private Load Balancer, Private Worker Nodes) setup, they can try and use bastion host to connect.