To avoid overwriting files, configure the File/Hadoop CSV Input and Output adapter to
only read and write files in the sandbox base directory. If the adapter is running in
unmanaged mode, update the controller.xml file to change the directory
in which the File/Hadoop CSV Input and Output adapter reads and writes files to the adapter
sandbox base directory.
Procedure
- In the controller.xml file, configure the following
parameters:
- For AdapterSandboxBase, set the path of the sandbox base
directory.
- For EnableSandboxing, set the value to true.
For
example,
<Controller>
<ControlPort>
<MinPort>19082</MinPort>
<MaxPort></MaxPort>
</ControlPort>
<Statistics_Update_Interval>5000</Statistics_Update_Interval>
<MonitorInterval>3000</MonitorInterval>
<AdapterSandboxBase><sanbox_base_directory_file_path></AdapterSandboxBase>
<EnableSandboxing>true</EnableSandboxing>
</Controller>
- Set the environment variable STREAMING_SANDBOX_ENABLE to
TRUE.
For the File/Hadoop CSV Input adapter, the file path configured in the File Input
Transporter will redirect all file access to the subfolder of the directory
configured in
AdapterSandboxBase. For the File/Hadoop CSV
Output adapter, the file path configured in the File Output Transporter will
redirect all file access to the subfolder of the directory configured in
AdapterSandboxBase.
Note For both the File/Hadoop CSV
Input and Output adapter, the sandbox base directory does not redicrect
files to and from a Hadoop Distributed File System (HDFS).