I tried to keep the same credentials to S3, but in the server side, the error points to the process doesn't have permissions to access the file. Does the server's user have to ask for reading privileges in the bucket?
Page 1 / 1
It depends a little on how you are accessing the file in the bucket.
This is a snippet from the FME Doc:
"
The S3Connector can use credentials from different sources. Using a web connection integrates best with FME, but in some cases, you may wish to use one of the other sources.
- Anonymous - access public S3 buckets without authenticating
- Web Connection - use an Amazon Web Services web connection stored in the FME web connections database
- Embedded - embed an access key ID and secret access key as parameters in the transformer
- System - use credentials already configured on the system. This allows for using IAM role-based authentication. For more information about configuring AWS credentials system-wide, see https://docs.aws.amazon.com/cli/latest/userguide/cli-chap-getting-started.html#config-settings-and-precedence
"
If you have set up a web connection in the workspace then you will need to make sure the web connection is published and the user running the worksapce has access to the web connection.
https://community.safe.com/s/article/fme-server-connections-runtime-error-unauthorized
If on the other hand you are using System then you will need to make sure that the FME Server system (and yeah I guess specifically the user running the FMEEngine service) will need to be configured.
Using embedded should work, however, it's also possible that there are additional IP range restrictions which might restrict access from only certain IP's. If the FME Server is outside of that environment then the server might need to get specific access.
It was about modifying an existing workspace, that used to take its csv files from a specific URL; currently the source is a S3 bucket. The solution was to build a new workspace that preserves the writers as the original.