Redshift Data Sharing - Play and Pause
Is there any change in "Play and Pause" functionality and costs for compute redshift clusters that are sharing data or using shared data in "Redshift data sharing feature"?
For example,
- Cluster A (RA3) shares data with cluster B (DC2) and Cluster C (RA3).
- Cluster A is only crunching data during off-office hours and is paused during business hours.
- Cluster A shares data (not all) with Cluster B and Cluster C
- Cluster B and C are paused during off-office hours and are active only during business hours
So, the question is how Cluster A, Cluster B and Cluster C are charged?
PS: I know I can use cluster A during business hours and save with RI but just assume here I cannot use cluster A because cluster A, B and C belong to different Line of Business.
- Newest
- Most votes
- Most comments
Hello Vikas, at reInvent 2020, Redshift released a feature called live data sharing across clusters. Currently this feature is in preview and works for the RA3 family of clusters. The producer cluster stores all the data and shares the desired data to the consumer clusters.
The producer cluster pays for the storage for the share (obviously) and the consumer will pay for the compute as and when it tries to query the share.
This is a live feature, which means as soon as there is a DDL / DML change in the producer cluster is reflected in the consumer right away. If the producer is down for some reason such as maintenance / paused then this feature will not work.
Have a look at the resources below for further reading.
Relevant content
- asked 2 years ago
- Accepted Answerasked 2 months ago
- asked 8 months ago
AWS OFFICIALUpdated a year ago- AWS OFFICIALUpdated a year ago
- AWS OFFICIALUpdated 2 years ago
- AWS OFFICIALUpdated a year ago
Is this still accurate, https://docs.aws.amazon.com/redshift/latest/dg/cluster-mgmt.html states that when a cluster is paused the share is still available and in fact other resources state the caching continues to be maintained.