INSTALLATION and GETTING STARTED

#       $Id: INSTALL,v 11.3 2003/08/29 09:55:03 jw Exp $  $Name:  $


PREREQUISITES

	rsync version 2.5.6 or higher.
	rsync version 2.5.7 or higher if backing up windows.

	perl5 and these perl modules (if you aren't familiar with
	perl see CPAN(3pm)).
		POSIX
		Getopt::Long
		Time::ParseDate
		Time::Period by Patric Ryan
	
	Lots of empty disk space.  I recommend using volume management
	(see lvm(8)).

	Ssh configured so root on the backup server can use ssh
	non-interactively to the backup clients.  This does not
	necessarily mean it has to be able to be root on those clients.

INSTALL

	Execute the install script "sh install.sh" from inside the
	unpack/download directory.  The script will ask where to
	install the various pieces.  Be sure you have write
	permissions for the installation directories.


CONFIGURATION

	Dirvish will store the backup images in "vault"s.  Plan on one
	vault per backup set.

	Create one or more directories as "bank"s in which to place
	the vaults.  You can create a large filesystem for each bank
	but it is usually better to create a filesystem for each
	vault.

	Inside the banks create one mount-point directory (vault) for
	each backup set (ex: home).  Start by making each filesystem
	about 1.5 times the size of the filesystem or directory tree
	it will be used to back up.  Mount 'em up.  The FAQ has
	additional details about how these filesystems should be
	built.
	
	create the file /etc/dirvish/master.conf on your backup server.
	It might look something like this:
		bank:
			/e/backup1
			/e/backup2
		exclude:
			lost+found/
			core
			*~
			.nfs*
		Runall:
			app	22:00
			data	22:00
			home	22:00
			site	22:00
		expire-default: +15 days
		expire-rule:
		#       MIN HR    DOM MON       DOW  STRFTIME_FMT
			*   *     *   *         1    +3 months
			*   *     1-7 *         1    +1 year
			*   *     1-7 1,4,7,10  1
			*   10-20 *   *         *    +4 days


	In each vault create a dirvish subdirectory.  In the dirvish
	subdirectory of each vault create config files.  To keep it
	simple here we will assume will only have one branch per vault
	and use the default configfile name; so create
	$BANK/$VAULT/dirvish/default.conf files.

	Here is an example from my home vault:
		client: leto
		tree: /e/home /home
		branch-default: daily
		exclude:
		        /**/.*/**/*cache/
			/**/.*/*cache/
			/**/.*/**/*Cache/
			.kde/share/thumbnails/

	Now that you are configured create an initial image by running
	dirvish on the backup server like so:

		root# dirvish --vault home --init
	
	If you add the --no-run option you will see what it will try
	to do.

	To do a regular backup you now only would need to execute
	"dirvish --vault home"

	After you create your image you can look at the log and summary
	files to see what was done.

	I strongly recommend reading the dirvish and dirvish.conf manpages.

	Work on your exclude lists and have fun.

	dirvish.runall will run dirvish on all of the vault[:branch]s
	listed in the Runall list in /etc/dirvish.conf

	Because dirvish handles calculating expiration dates
	automatically you need only add two entries to crontab to
	completely automate your dirvish backups.

		10 03 * * *     root /site/sbin/dirvish-expire --quiet
		18 03 * * *     root /site/sbin/dirvish-runall --quiet

	Note that both dirvish-runall and dirvish-expire are scheduled
	after 22:00 as specified in the Runall: parameter.