aws-redshift

TLDR

List all clusters

$ aws redshift describe-clusters

Describe a specific cluster

$ aws redshift describe-clusters --cluster-identifier [cluster_name]

Create a new cluster

$ aws redshift create-cluster --cluster-identifier [cluster_name] --node-type dc2.large --master-username [admin] --master-user-password [password] --number-of-nodes [2]

Create a single-node cluster

$ aws redshift create-cluster --cluster-identifier [cluster_name] --node-type dc2.large --master-username [admin] --master-user-password [password] --cluster-type single-node

Resize a cluster

$ aws redshift resize-cluster --cluster-identifier [cluster_name] --cluster-type multi-node --node-type ra3.xlplus --number-of-nodes [4]

Create a snapshot

$ aws redshift create-cluster-snapshot --cluster-identifier [cluster_name] --snapshot-identifier [snapshot_name]

Delete a cluster (skip final snapshot)

$ aws redshift delete-cluster --cluster-identifier [cluster_name] --skip-final-cluster-snapshot

Pause a cluster to save costs

$ aws redshift pause-cluster --cluster-identifier [cluster_name]

SYNOPSIS

aws redshift subcommand [options]

aws redshift is a subcommand of the AWS CLI that manages Amazon Redshift, a fully managed petabyte-scale data warehouse service. Redshift uses columnar storage and parallel query execution for fast analytics on large datasets.
Clusters consist of leader and compute nodes. The leader node manages connections and query planning, while compute nodes store data and execute queries. Redshift supports SQL queries through standard PostgreSQL drivers.
Redshift Serverless provides a pay-per-use option without managing clusters. Data can be loaded from S3, DynamoDB, EMR, or streamed via Firehose.

PARAMETERS

describe-clusters

List clusters and their details.

create-cluster

Launch a new cluster.

delete-cluster

Terminate a cluster.

modify-cluster

Change cluster settings.

resize-cluster

Change node type or count.

pause-cluster

Pause a cluster to stop billing.

resume-cluster

Resume a paused cluster.

reboot-cluster

Restart a cluster.

create-cluster-snapshot

Create a manual snapshot.

restore-from-cluster-snapshot

Create cluster from snapshot.

describe-cluster-snapshots

List snapshots.

--cluster-identifier name

Unique cluster name.

--node-type type

Instance type (dc2.large, ra3.xlplus, etc.).

--number-of-nodes count

Number of compute nodes.

--cluster-type type

single-node or multi-node.

--master-username name

Admin username.

--master-user-password password

Admin password.

--db-name name

Default database name.

CAVEATS

Cluster creation takes several minutes. Classic resize creates a new cluster and migrates data; elastic resize is faster but has limitations. Paused clusters still incur storage costs. Multi-node clusters require minimum of 2 nodes. Master password must meet complexity requirements. Public access requires proper security group and VPC settings.

HISTORY

Amazon Redshift launched in February 2013 as AWS's first data warehouse service, named after the astronomical phenomenon. Redshift Spectrum was introduced in 2017 for querying data directly in S3. RA3 nodes with managed storage launched in 2019, separating compute and storage. Redshift Serverless became generally available in July 2022. The service has added features like AQUA (Advanced Query Accelerator), ML integration, and data sharing.