Redshift data share error

0

Hi

We are using datashare to share data between 2 redshift clusters within the same account. We started running into this started recently, no information regarding this on web or in aws docs.

Any inputs are appreciate on how to solve this.

ERROR: ----------------------------------------------- error: Data Sharing Internal error. code: 34519 context: query: 0 location: burst_refresh_controller.cpp:481 process: padbmaster [pid=1073997230] ----------------------------------------------- [ErrorId: 1-64d11743-5caa5b6723a0e8266f34b311]

  • Few things to try:

    Make sure both clusters are in the same AWS account and region. Cross-account and cross-region data shares can cause issues. Check that the IAM roles and permissions are set up correctly between the clusters. The role on the producer cluster needs access to the consumer cluster. Try recreating the data share or set up a new one between the clusters. There may be some metadata corruption on the existing one. Check your VPC and security group configurations on both clusters. There needs to be connectivity and no blocking of ports. Look at the Redshift logs on both clusters for any other related errors or warnings. Try a lower load to rule out any issues with volume - share fewer tables or limit sync frequency.

  • Try recreating the data share or set up a new one between the clusters. There may be some metadata corruption on the existing one.

    Creating a new datashare seem to run into the same issue as well

asked 10 months ago641 views
1 Answer
0

Review the IAM roles and permissions in detail - ensure the data share role has the redshift:DescribeClusters and redshift: DescribeClusterSnapshots permissions required. Check S3 bucket access - the role needs to Get access to the bucket where data is staged. Verify the security groups allow all traffic between the Redshift clusters on the data share port 5439. Check if both clusters are in the same VPC and availability zone. Cross AZ shares may have issues. Try creating a share between a different producer and consumer cluster. This can help isolate if a specific cluster is having problems. Scale down the data being shared to the minimum - a single table, no incrementals. See if a very small share works. Capture VPC flow logs during the data share to check for any rejected connections

answered 10 months ago

You are not logged in. Log in to post an answer.

A good answer clearly answers the question and provides constructive feedback and encourages professional growth in the question asker.

Guidelines for Answering Questions