Transferring files between Amazon RDS for Oracle and an Amazon S3 bucket

To transfer files between an RDS for Oracle DB instance and an Amazon S3 bucket, you can use the Amazon RDS package rdsadmin_s3_tasks. You can compress files with GZIP when uploading them, and decompress them when downloading.

Requirements and limitations for file transfers

Before transferring files between your DB instance and an Amazon S3 bucket, note the following:

• The rdsadmin_s3_tasks package transfers files located in a single directory. You can't include subdirectories in a transfer.

• The maximum object size in an Amazon S3 bucket is 5 TB.

• Tasks created by rdsadmin_s3_tasks run asynchronously.

• You can upload files from the Data Pump directory, such as DATA_PUMP_DIR, or any user-created directory. You can't upload files from a directory used by Oracle background processes, such as the adump, bdump, or trace directories.

• The download limit is 2000 files per procedure call for download_from_s3. If you need to download more than 2000 files from Amazon S3, split your download into separate actions, with no more than 2000 files per procedure call.

• If a file exists in your download folder, and you attempt to download a file with the same name, download_from_s3 skips the download. To remove a file from the download directory, use the PL/SQL procedure UTL_FILE.FREMOVE.

Uploading files from your RDS for Oracle DB instance to an Amazon S3 bucket

To upload files from your DB instance to an Amazon S3 bucket, use the procedure rdsadmin.rdsadmin_s3_tasks.upload_to_s3. For example, you can upload Oracle Recovery Manager (RMAN) backup files or Oracle Data Pump files. For more information about working with objects, see Amazon Simple Storage Service User Guide. For more information about performing RMAN backups, see Performing common RMAN tasks for Oracle DB instances.

The rdsadmin.rdsadmin_s3_tasks.upload_to_s3 procedure has the following parameters.

Parameter name	Data type	Default	Required	Description
p_bucket_name	VARCHAR2	–	required	The name of the Amazon S3 bucket to upload files to.
p_directory_name	VARCHAR2	–	required	The name of the Oracle directory object to upload files from. The directory can be any user-created directory object or the Data Pump directory, such as DATA_PUMP_DIR. You can't upload files from a directory used by background processes, such as adump, bdump, and trace. Note You can only upload files from the specified directory. You can't upload files in subdirectories in the specified directory.
p_s3_prefix	VARCHAR2	–	required	An Amazon S3 file name prefix that files are uploaded to. An empty prefix uploads all files to the top level in the specified Amazon S3 bucket and doesn't add a prefix to the file names. For example, if the prefix is folder_1/oradb, files are uploaded to folder_1. In this case, the oradb prefix is added to each file.
p_prefix	VARCHAR2	–	required	A file name prefix that file names must match to be uploaded. An empty prefix uploads all files in the specified directory.
p_compression_level	NUMBER	0	optional	The level of GZIP compression. Valid values range from 0 to 9: 0 – No compression 1 – Fastest compression 9 – Highest compression
p_bucket_owner_full_control	VARCHAR2	–	optional	The access control setting for the bucket. The only valid values are null or FULL_CONTROL. This setting is required only if you upload files from one account (account A) into a bucket owned by a different account (account B), and account B needs full control of the files.

The return value for the rdsadmin.rdsadmin_s3_tasks.upload_to_s3 procedure is a task ID.

The following example uploads all of the files in the DATA_PUMP_DIR directory to the Amazon S3 bucket named amzn-s3-demo-bucket. The files aren't compressed.

SELECT rdsadmin.rdsadmin_s3_tasks.upload_to_s3(
      p_bucket_name    =>  'amzn-s3-demo-bucket',
      p_prefix         =>  '', 
      p_s3_prefix      =>  '', 
      p_directory_name =>  'DATA_PUMP_DIR') 
   AS TASK_ID FROM DUAL;

The following example uploads all of the files with the prefix db in the DATA_PUMP_DIR directory to the Amazon S3 bucket named amzn-s3-demo-bucket. Amazon RDS applies the highest level of GZIP compression to the files.

SELECT rdsadmin.rdsadmin_s3_tasks.upload_to_s3(
      p_bucket_name       =>  'amzn-s3-demo-bucket', 
      p_prefix            =>  'db', 
      p_s3_prefix         =>  '', 
      p_directory_name    =>  'DATA_PUMP_DIR',
      p_compression_level =>  9) 
   AS TASK_ID FROM DUAL;

The following example uploads all of the files in the DATA_PUMP_DIR directory to the Amazon S3 bucket named amzn-s3-demo-bucket. The files are uploaded to a dbfiles folder. In this example, the GZIP compression level is 1, which is the fastest level of compression.

SELECT rdsadmin.rdsadmin_s3_tasks.upload_to_s3(
      p_bucket_name       =>  'amzn-s3-demo-bucket', 
      p_prefix            =>  '', 
      p_s3_prefix         =>  'dbfiles/', 
      p_directory_name    =>  'DATA_PUMP_DIR',
      p_compression_level =>  1) 
   AS TASK_ID FROM DUAL;        

The following example uploads all of the files in the DATA_PUMP_DIR directory to the Amazon S3 bucket named amzn-s3-demo-bucket. The files are uploaded to a dbfiles folder and ora is added to the beginning of each file name. No compression is applied.

SELECT rdsadmin.rdsadmin_s3_tasks.upload_to_s3(
      p_bucket_name    =>  'amzn-s3-demo-bucket', 
      p_prefix         =>  '', 
      p_s3_prefix      =>  'dbfiles/ora', 
      p_directory_name =>  'DATA_PUMP_DIR') 
   AS TASK_ID FROM DUAL;

The following example assumes that the command is run in account A, but account B requires full control of the bucket contents. The command rdsadmin_s3_tasks.upload_to_s3 transfers all files in the DATA_PUMP_DIR directory to the bucket named s3bucketOwnedByAccountB. Access control is set to FULL_CONTROL so that account B can access the files in the bucket. The GZIP compression level is 6, which balances speed and file size.

SELECT rdsadmin.rdsadmin_s3_tasks.upload_to_s3(
      p_bucket_name               =>  's3bucketOwnedByAccountB', 
      p_prefix                    =>  '', 
      p_s3_prefix                 =>  '', 
      p_directory_name            =>  'DATA_PUMP_DIR',
      p_bucket_owner_full_control =>  'FULL_CONTROL',
      p_compression_level         =>  6) 
   AS TASK_ID FROM DUAL;

In each example, the SELECT statement returns the ID of the task in a VARCHAR2 data type.

You can view the result by displaying the task's output file.

SELECT text FROM table(rdsadmin.rds_file_util.read_text_file('BDUMP','dbtask-task-id.log'));

Replace task-id with the task ID returned by the procedure.

Note

Tasks are executed asynchronously.

Downloading files from an Amazon S3 bucket to an Oracle DB instance

To download files from an Amazon S3 bucket to an RDS for Oracle instance, use the Amazon RDS procedure rdsadmin.rdsadmin_s3_tasks.download_from_s3.

The download_from_s3 procedure has the following parameters.

Parameter name	Data type	Default	Required	Description
p_bucket_name	VARCHAR2	–	Required	The name of the Amazon S3 bucket to download files from.
p_directory_name	VARCHAR2	–	Required	The name of the Oracle directory object to download files to. The directory can be any user-created directory object or the Data Pump directory, such as DATA_PUMP_DIR.
p_error_on_zero_downloads	VARCHAR2	FALSE	Optional	A flag that determines whether the task raises an error when no objects in the Amazon S3 bucket match the prefix. If this parameter is not set or set to FALSE (default), the task prints a message that no objects were found, but doesn't raise an exception or fail. If this parameter is TRUE, the task raises an exception and fails. Examples of prefix specifications that can fail match tests are spaces in prefixes, as in ' import/test9.log', and case mismatches, as in test9.log and test9.LOG.
p_s3_prefix	VARCHAR2	–	Required	A file name prefix that file names must match to be downloaded. An empty prefix downloads all of the top level files in the specified Amazon S3 bucket, but not the files in folders in the bucket. The procedure downloads Amazon S3 objects only from the first level folder that matches the prefix. Nested directory structures matching the specified prefix are not downloaded. For example, suppose that an Amazon S3 bucket has the folder structure folder_1/folder_2/folder_3. You specify the 'folder_1/folder_2/' prefix. In this case, only the files in folder_2 are downloaded, not the files in folder_1 or folder_3. If, instead, you specify the 'folder_1/folder_2' prefix, all files in folder_1 that match the 'folder_2' prefix are downloaded, and no files in folder_2 are downloaded.
p_decompression_format	VARCHAR2	–	Optional	The decompression format. Valid values are NONE for no decompression and GZIP for decompression.

The return value for the rdsadmin.rdsadmin_s3_tasks.download_from_s3 procedure is a task ID.

The following example downloads all files in the Amazon S3 bucket named amzn-s3-demo-bucket to the DATA_PUMP_DIR directory. The files aren't compressed, so no decompression is applied.

SELECT rdsadmin.rdsadmin_s3_tasks.download_from_s3(
      p_bucket_name    =>  'amzn-s3-demo-bucket',
      p_directory_name =>  'DATA_PUMP_DIR') 
   AS TASK_ID FROM DUAL;

The following example downloads all of the files with the prefix db in the Amazon S3 bucket named amzn-s3-demo-bucket to the DATA_PUMP_DIR directory. The files are compressed with GZIP, so decompression is applied. The parameter p_error_on_zero_downloads turns on prefix error checking, so if the prefix doesn't match any files in the bucket, the task raises and exception and fails.

SELECT rdsadmin.rdsadmin_s3_tasks.download_from_s3(
      p_bucket_name               =>  'amzn-s3-demo-bucket', 
      p_s3_prefix                 =>  'db', 
      p_directory_name            =>  'DATA_PUMP_DIR',
      p_decompression_format      =>  'GZIP',
      p_error_on_zero_downloads   =>  'TRUE') 
   AS TASK_ID FROM DUAL;

The following example downloads all of the files in the folder myfolder/ in the Amazon S3 bucket named amzn-s3-demo-bucket to the DATA_PUMP_DIR directory. Use the p_s3_prefix parameter to specify the Amazon S3 folder. The uploaded files are compressed with GZIP, but aren't decompressed during the download.

SELECT rdsadmin.rdsadmin_s3_tasks.download_from_s3(
      p_bucket_name          =>  'amzn-s3-demo-bucket', 
      p_s3_prefix            =>  'myfolder/', 
      p_directory_name       =>  'DATA_PUMP_DIR',
      p_decompression_format =>  'NONE')
   AS TASK_ID FROM DUAL;

The following example downloads the file mydumpfile.dmp in the Amazon S3 bucket named amzn-s3-demo-bucket to the DATA_PUMP_DIR directory. No decompression is applied.

SELECT rdsadmin.rdsadmin_s3_tasks.download_from_s3(
      p_bucket_name    =>  'amzn-s3-demo-bucket', 
      p_s3_prefix      =>  'mydumpfile.dmp', 
      p_directory_name =>  'DATA_PUMP_DIR') 
   AS TASK_ID FROM DUAL;

In each example, the SELECT statement returns the ID of the task in a VARCHAR2 data type.

You can view the result by displaying the task's output file.

SELECT text FROM table(rdsadmin.rds_file_util.read_text_file('BDUMP','dbtask-task-id.log'));

Replace task-id with the task ID returned by the procedure.

Note

Tasks are executed asynchronously.

You can use the UTL_FILE.FREMOVE Oracle procedure to remove files from a directory. For more information, see FREMOVE procedure in the Oracle documentation.

Monitoring the status of a file transfer

File transfer tasks publish Amazon RDS events when they start and when they complete. The event message contains the task ID for the file transfer. For information about viewing events, see Viewing Amazon RDS events.

You can view the status of an ongoing task in a bdump file. The bdump files are located in the /rdsdbdata/log/trace directory. Each bdump file name is in the following format.

dbtask-task-id.log

Replace task-id with the ID of the task that you want to monitor.

Note

Tasks are executed asynchronously.

You can use the rdsadmin.rds_file_util.read_text_file stored procedure to view the contents of bdump files. For example, the following query returns the contents of the dbtask-1234567890123-1234.log bdump file.

SELECT text FROM table(rdsadmin.rds_file_util.read_text_file('BDUMP','dbtask-1234567890123-1234.log'));

The following sample shows the log file for a failed transfer.

TASK_ID                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                         
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
1234567890123-1234


TEXT                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                            
--------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------
2023-04-17 18:21:33.993 UTC [INFO ] File #1: Uploading the file /rdsdbdata/datapump/A123B4CDEF567890G1234567890H1234/sample.dmp to Amazon S3 with bucket name amzn-s3-demo-bucket and key sample.dmp.
2023-04-17 18:21:34.188 UTC [ERROR] RDS doesn't have permission to write to Amazon S3 bucket name amzn-s3-demo-bucket and key sample.dmp.
2023-04-17 18:21:34.189 UTC [INFO ] The task failed.