[ 
https://issues.apache.org/jira/browse/CASSANALYTICS-155?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18081002#comment-18081002
 ] 

Jon Haddad commented on CASSANALYTICS-155:
------------------------------------------

I've verified this works in this [easy-db-lab 
branch|https://github.com/rustyrazorblade/easy-db-lab/tree/31ec4c8f4e040e0a0da257b7bdf4e564bc64284f/spark/bulk-writer-s3-iam].

[Storage 
Extension|https://github.com/rustyrazorblade/easy-db-lab/blob/31ec4c8f4e040e0a0da257b7bdf4e564bc64284f/spark/bulk-writer-s3-iam/src/main/java/com/rustyrazorblade/easydblab/spark/EasyDbLabIamStorageExtension.java]

[Script to run it 
all.|https://github.com/rustyrazorblade/easy-db-lab/blob/31ec4c8f4e040e0a0da257b7bdf4e564bc64284f/bin/test-spark-bulk-writer-s3-iam]

Relevant output from the spark job:
 
{noformat}
=== Submitting IAM bulk writer Spark job ===
Uploading bulk-writer-s3-iam-all.jar to 
s3://easy-db-lab-data-6afa29e6-338a-47eb-95e2-568ec44a3792/spark/bulk-writer-s3-iam-all.jar...
Upload complete: 
s3://easy-db-lab-data-6afa29e6-338a-47eb-95e2-568ec44a3792/spark/bulk-writer-s3-iam-all.jar
Submitted Spark job: s-04517261ITV6J5DMUKN9 to cluster j-UEAA3KYPRF1
Waiting for job completion...
Job state: RUNNING
Job completed successfully
=== Verifying row count ===
"count"
----------
1000000
=== Done ===
{noformat}
 

 

> support IAM when bulk writing to S3
> -----------------------------------
>
>                 Key: CASSANALYTICS-155
>                 URL: https://issues.apache.org/jira/browse/CASSANALYTICS-155
>             Project: Apache Cassandra Analytics
>          Issue Type: Improvement
>            Reporter: Jon Haddad
>            Priority: Normal
>
> When triggering a restore job via Apache Cassandra Sidecar, the analytics 
> library currently requires explicit static S3 credentials to be passed in the 
> job request. This prevents deployments where the Sidecar node authenticates 
> to S3 via IAM instance profiles or other non-static credential providers.
> CASSSIDECAR-415 adds support on the Sidecar side for a credentialType field 
> in the restore job request, allowing the Sidecar to use its own IAM role when 
> credentials are not provided.
> This ticket tracks the corresponding changes in Cassandra Analytics to allow 
> users to configure a CredentialType on StorageCredentialPair, selecting how 
> the Sidecar should authenticate to S3. The initial supported types are STATIC 
> (existing behavior, no change) and IAM. The design leaves room for additional 
> credential types in the future without structural changes.
> All existing restore job functionality (coordinated writes, consistency 
> levels, cross-region support, lease management) is unaffected — only the 
> authentication method passed to the Sidecar changes.
> The same authentication method passed to sidecar will also need to be used to 
> push the files to S3.



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to