rsync

Introduction
- A File Transfer Protocol
- Optimised for the case where similar files are already on the
  receiving end
- Protocol
- Usage
  - Transports
  - Switches
- Applications
- Other things

The rsync Protocol
- File going from A to B
- diagram showing how the data is broken up
- weak fast rolling checksum
- strong hash (MD4)
- strong hash on whole file

Transports rsync runs over:

- local
  rsync src dest

- rsh
  - rsync user@host:src dest
  - rsync src user@host:dest
  - requires a login on the remote machine

- ssh
  - rsync -e ssh user@host:src dest
  - rsync -e ssh src user@host:dest
  - export RSYNC_RSH=ssh
  - requires a login on the remote machine

- rsync
  - rsync user@host::src dest
  - rsync src user@host::src
  - may or may not require a login (can be anonymous)
  - rsyncd (run as either daemon or inetd)
  - /etc/rsyncd.conf:

    [tivo]
      comment = Tivo Stuff
      path = /h/g/geoffrey/tivo
      uid = g
      read only = no
      hosts allow = 10.0.0.0/24 127.0.0.1

Common Switches:
- -a (archive) -- recursive, copy all attributes
- -v (verbose)
- -z (compress)
- --exclude
- --delete
- --dry-run
- --partial
- --progress
- -P (equivalent to --partial and --progress)

Applications:
- FTP and scp replacement
- backup
  - to a local disk or over a network
- mirroring
  - FTP servers
  - web servers
  - databases
  - files->ISO example
- some FTP servers offer rsync access to the public (eg.
  mirror.aarnet.edu.au)

Other optimisations
- file list transfered with rsync
- pipelining

Block size discussion

Problems

- compressed files
- moved files