Editing Rsync (section)

== Connection ==
An rsync process operates by communicating with another rsync process, a sender and a receiver. At startup, an rsync client connects to a peer process. If the transfer is local (that is, between file systems mounted on the same host) the peer can be created with fork, after setting up suitable pipes for the connection. If a remote host is involved, rsync starts a process to handle the connection, typically [[Secure Shell]]. Upon connection, a command is issued to start an rsync process on the remote host, which uses the connection thus established. As an alternative, if the remote host runs an rsync daemon, rsync clients can connect by opening a socket on TCP port 873, possibly using a proxy.<ref>{{Cite web |title=How Rsync Works |url=https://rsync.samba.org/how-rsync-works.html |url-status=live |archive-url=https://web.archive.org/web/20161216173537/http://rsync.samba.org/how-rsync-works.html |archive-date=16 December 2016 |access-date=24 January 2017}}</ref>

Rsync has numerous command line options and configuration files to specify alternative shells, options, commands, possibly with full path, and port numbers.  Besides using remote shells, tunnelling can be used to have remote ports appear as local on the server where an rsync daemon runs. Those possibilities allow adjusting security levels to the state of the art, while a naive rsync daemon can be enough for a local network.

One solution is the <code>[[Dry run (testing)|--dry-run]]</code> option, which allows users to validate their [[command-line argument]]s and to simulate what would happen when copying the data without actually making any changes or transferring any data.