-
Notifications
You must be signed in to change notification settings - Fork 302
Databricks
Yitao Li edited this page Jun 22, 2020
·
1 revision
- Deployment was mostly straightforward. Just make sure to choose the EC2 region closest to you. Otherwise there may be large latency with SSH connections later on.
- Make sure to include your public key in the SSH tab before clicking "create cluster" so that the nodes could be accessed via SSH later on
- There is also a way to inject authorized SSH key using Databricks notebook after cluster deployment, as documented in https://docs.databricks.com/clusters/configure.html#configure-an-existing-cluster-with-your-public-key
-
Must use port 2200 (not 22) to SSH into Spark driver --
ssh ubuntu@<hostname> -p 2200 -i <private-key-file-path>
so the relevant security group needs to be changed to allow traffic from a given external IP on port 2200