![]() |
![]()
| ![]() |
![]()
NAME
DESCRIPTIONThe The In this document, "client" refers to the openrsync(1) utility making the request. It follows that "server" refers to the daemon servicing the request. A connection between a client and server consists of host, a module, and zero or more paths. openrsync rsync://host/module/path1
rsync://host/path2... dest At this time, operating in sender mode (with the rsync:// host receiving information) is not described in this document. Data typesThese are the same as in openrsync(5). A newline is always a standalone \n. Client processAfter initialising a connection, the client and server exchange the following information, in order. This portion of the process is not multiplexed.
The requested module must have non-zero length. The preamble consists in a sequence of lines. Each line either contains free-form text sent by the server as a "motd" (message of the day) or a command: @RSYNCD: command\n The only supported command is the server protocol specification: @RSYNCD: xx[.yy]\n The optional component is the submodule, which may be discarded. The version may only be specified once. Both the motd and commands end in the special termination command: @RSYNCD: OK\n Following that, the client must send the command-line arguments that would otherwise be used to start a openrsync(1) server. Each argument must be specified on its own line, e.g., --server\n--sender\n-r\n-t\n.\npath1\npath2 This must be followed by a standalone newline. If the server does not understand or accept any of the command-line arguments, it will exit at this point. Following this, the client must read the integer-length session checksum seed. Multiplexing is subsequently enabled. The sequence that follows is stipulated in openrsync(5) following the handshake. SEE ALSO
|