Introduction

One of the most efficient ways to load data from Amazon Redshift to s3 is using the UNLOAD command:  http://docs.aws.amazon.com/redshift/latest/dg/t_Unloading_tables.html

Use case(s)

• A financial customer would like to quickly unload financial reports into S3 that have been generated from processing that is happening in Redshift. The pipeline would have a Redshift to S3 action at the beginning, and then leverage the s3 source to read that data into a processing pipeline.

User Storie(s)

• As a user, i would like to unload data from redshift to s3 using the unload command.
• I would like to authenticate with IAM credentials as well as id and secret key pairs.
• I would like to use that s3 data as an input into a hydrator pipeline
• I would like the location of the data to be passed via workflow token so that they next plugin can use it in a macro

Plugin Type

• Action

Configurables

This section defines properties that are configurable for this plugin. 

Design / Implementation Tips

• Follow the tutorial here to get a better understanding of what might be http://docs.aws.amazon.com/redshift/latest/dg/t_Unloading_tables.html

Design

1. User can connect to the S3 buckets using Either access and secret access key or using IAM role.
2. For using the IAM role to unload the data, the IAM role should have GET, LIST and PUT access on the bucket.
3. Existing S3 source plugins read data with regions that have versions 2 and 4 signature version support. Hence, RedhsiftToS3 plugin will have support for only these regions: "us-east-1", "us-west-1", "us-west-2", "ap-southeast-1", "ap-southeast-2", "ap-northeast-1", "eu-west-1" and "sa-east-1".
4. UNLOAD automatically creates files using Amazon S3 server-side encryption with AWS-managed encryption keys (SSE-S3). UNLOAD does not support Amazon S3 server-side encryption with encryption keys from SSE-KMS or a customer-supplied key (SSE-C).
5. Plugin would always unload data with Amazon S3 server-side encryption.
6. The user should provide complete JDBC Redshift DB url for connecting to the redshift cluster. Any error in the connection string will result in run-time exception.
7. If the table mentioned in the select query in unload, does not exists, the plugin will throw run-time exception.
8.  If the user passes name-prefix in the S3 data path, the plugin will append * to the path and pass this new location as a workflow token.
   The next stage will then use file path globbing to read all unloaded files starting with the prefix.

Approach(s)

1. Plugin will check if the query starts with select and contains the from clause. Any other error, in the SQL statement will result in run-time failure.
2. The user can either provide the credentials : accessKey and secretAccessKey or the IAM role (which has GET, LIST and PUT permissions on the bucket) for unloading the data from redshift to S3.
3. S3 bucket is the full path, including bucket name, to the location on Amazon S3 where Amazon Redshift will write the output file objects, including the manifest file if MANIFEST is specified. The object names are prefixed with name-prefix. 
4. Single ASCII character that is used to separate fields in the output file, such as a pipe character ( | ), a comma ( , ), or a tab ( \t ), can be specified using the config parameter "delimiter". The default delimiter is a pipe character. If the data contains the delimiter character, user will need to specify the ESCAPE option to escape the delimiter, or use ADDQUOTES to enclose the data in double quotes.

Properties

{
  "name": "RedshiftToS3Action",
  "type": "action",
  "properties": {
    "accessKey": "accessKey",
    "secretAccessKey": "secretAccessKey",
    "iamRole": "",
    "query": "select * from venue",
    "s3DataPath": "s3://mybucket/test_",
    "outputPathToken": "",
    "manifest": "false",
    "delimiter": ",",
    "parallel": "false",
    "compression" : "NONE",
    "allowOverwrite" : "true",
    "addQuotes": "true",
    "escape": "false",
    "redshiftClusterURL" : "jdbc:redshift://test.cntuu3e3qg5h.us-west-1.redshift.amazonaws.com:5439/redshiftdb",
    "redshiftMasterUser" : "admin",
    "redshiftMasterPassword": "admin"
  }
}

Security

Limitation(s)

1. The Amazon S3 bucket where Amazon Redshift will write the output files must reside in the same region as your cluster.(http://docs.aws.amazon.com/redshift/latest/dg/r_UNLOAD.html)

Future Work

• Some future work – HYDRATOR-99999
• Another future work – HYDRATOR-99999

Test Case(s)

• RedshiftToS3 action - Using keys for connection

• RedshiftToS3 action - Using IAM Role for connection

• RedshiftToS3 action - Using compression option GZIP
• RedshiftToS3 action - Using parallel true

Sample Pipeline

RedshiftToS3- with Keys

RedshiftToS3- with IAM Role

RedshiftToS3 - with Compression type GZIP

RedshiftToS3- with Parallel true