If an upload to Amazon S3 fails, by default, `Uploader` uses the Amazon S3 `AbortMultipartUpload` operation to remove the uploaded parts. This functionality ensures that failed uploads do not consume Amazon S3 storage.
You can set `LeavePartsOnError` to true so that the `Uploader` doesn’t delete successfully uploaded parts. This is useful for resuming partially completed uploads. To operate on uploaded parts, you must get the `UploadID` of the failed upload. The following example demonstrates how to use the `manager.MultiUploadFailure` error interface type to get the `UploadID`.