Hi,
On 7/20 we upgraded from 22.1 to 22.2. We are using the Amazon S3 Utils version 1.1 and we can no longer upload to S3 after the upgrade.
This plug-in [com.appiancorp.ps.plugins.amazonS3Utils] is not registered to access secured values for the given external system key [s3prod]. Check the external systems plug-ins list in the Administration Console.
Error Occured: Error while uploading the file to Amazon S3. Error is: The security token included in the request is invalid. (Service: AWSKMS; Status Code: 400; Error Code: UnrecognizedClientException; Request ID: c1e7278c-ca73-46bd-bcf9-c510a620018a)
Thanks,
Scott
Discussion posts and replies are publicly visible
Have you had the chance to look at the std out log? Are there any more details in there? Also, was this tested and working in a lower environment after upgrade?
Here is another log entry we have noticed.
2022-07-25 16:07:43,697 [Appian Work Item - 8327 - WorkID 811 - execution01 - process 270508345 - model 155 : UnattendedJavaActivityRequest] ERROR com.appiancorp.ps.plugins.amazonS3Utils.UploadObjectsToAmazonS3 - Error while uploading the file to Amazon S3.com.amazonaws.AmazonServiceException: User: arn:aws:iam::124294179439:user/SCP is not authorized to perform: kms:GenerateDataKey on resource: arn:aws:kms:us-east-1:124294179439:key/0484e365-406c-4413-a453-781f530957f8 with an explicit deny in an identity-based policy (Service: AWSKMS; Status Code: 400; Error Code: AccessDeniedException; Request ID: 62e8514c-07a3-4602-b69d-af6243236224) at com.amazonaws.http.AmazonHttpClient.handleErrorResponse(AmazonHttpClient.java:1275) at com.amazonaws.http.AmazonHttpClient.executeOneRequest(AmazonHttpClient.java:873)
Looking at these errors, I do not think this is a plugin issue due to an upgrade. The second error easily points to a permissions set up on S3 - you need to check the authorization for account SCP. The first one is a bit more tricky but also likely due to an account issue, it could be that the key expired, or maybe even the account itself. That's where I would check first.
Thanks Mike. We are starting to look at other places that might be causing the issue.