Node Director
File Retrievers

File access drivers and URI documentation. More...


 Atomic File Retriever
 Local file access with files written in an atomic operation.
 Backup File Retriever
 Access local files and save a backup copy before overwriting.
 Config File Retriever
 Retrieve files and expand include statements.
 File Insert Retriever
 Access local files and preserve static sections within the file.
 File Retriever
 Access local files.
 Pipe Retriever
 Pipe files into shell commands or read them from commands.
 Server File Retriever
 Access files on the Director server via an agent connection.
 StdIO Retriever
 Read and write files via standard input/output.

Detailed Description

File access drivers and URI documentation.




pipe:dd of=/dev/null


A File Retriever is a driver implementing a way to access (read/write/truncate/append) a file. Generally, a file retriever implements some file access protocol. In the most simple case this might just be a file in the systems file system that can be accessed via ordinary file operations, but it can also be a remote file transmitted via FTP or HTTP.

Wherever the Director expects a file retriever URI (for instance as part of a database URI) a string of the form drivername:driverspec can be used, where drivername specifies which file access driver should be used and driverspec is a driver specific string that tells the driver which of the available files it should connect to.

Not every driver does necessarily provide all basic file I/O operations, there are files and protocols that only allow read-only or even write-only access, some cannot append to files but require a file to be transmitted as a whole (truncate+append).