#!/bin/bash
rsync –avp –delete /var /media/backup/local
rsync –avp –delete /etc /media/backup/local
rsync –avp — delete /home /media/backup/local

to run the script become root (as you’ll need that permission for /var and /etc) and run ‘./sys_backup’ without the quotes.
however, i wanted something a bit more comprehensive (and incremental) without me putting in too much effort. looked at bacula. too complicated. settled on backupninja. to install on a debian based system, as root, issue the following ‘apt-get install backupninja’ to grab it from the repos. next we want to edit the main configuration file. ‘joe /etc/backupninja.conf’ will open it and you can change the parameters to suit your setup. mine’s only a local backup so it looks a bit like this.
#                          |\_
# B A C K U P N I N J A   /()/
#                         ‘\|
# main configuration file
#

# how verbose to make the logs
# 5 — Debugging messages     (and below)
# 4 — Informational messages (and below)
# 3 — Warnings               (and below)
# 2 — Errors                 (and below)
# 1 — Fatal errors           (only)
loglevel = 4

# send a summary of the backup status to
# this email address:
reportemail = col@localhost

# if set to ‘yes’, a report email will be generated
# even if all modules reported success. (default = yes)
reportsuccess = yes

# if set to ‘yes’, info messages from handlers will be
# sent into the email (default = no)
reportinfo = no

# if set to ‘yes’, a report email will be generated
# even if there was no error. (default = yes)
reportwarning = yes

# if set to ‘yes’, disk space usage will be included in
# the backup email report
reportspace = yes

# where to rsync the backupninja.log to be aggregated in
# a ninjareport
reporthost =

# what user to connect to reporthost to sync the
# backupninja.log
reportuser = ninja

# where on the reporthost should the report go
# NOTE: the name of the log will be used in the report,
# use a globally unique name, preferably the hostname
reportdirectory = /var/lib/backupninja/reports

# set to the administration group that is allowed to
# read/write configuration files in /etc/backup.d
admingroup = root

#######################################################
# for most installations, the defaults below are good #
#######################################################

# where to log:
logfile = /var/log/backupninja.log

# directory where all the backup configuration files live
configdirectory = /etc/backup.d

# where backupninja helper scripts are found
scriptdirectory = /usr/share/backupninja

# where backupninja libs are found
libdirectory = /usr/lib/backupninja

# whether to use colors in the log file
usecolors = yes

# default value for ‘when’
when = everyday at 01:00

# if running vservers, set to yes
vservers = no

# programs paths
# SLAPCAT=/usr/sbin/slapcat
# LDAPSEARCH=/usr/bin/ldapsearch
# RDIFFBACKUP=/usr/bin/rdiff-backup
# CSTREAM=/usr/bin/cstream
# MYSQL=/usr/bin/mysql
# MYSQLHOTCOPY=/usr/bin/mysqlhotcopy
# MYSQLDUMP=/usr/bin/mysqldump
# PSQL=/usr/bin/psql
# PGSQLDUMP=/usr/bin/pg_dump
# PGSQLDUMPALL=/usr/bin/pg_dumpall
# GZIP=/bin/gzip
# GZIP_OPTS=’–rsyncable’
# RSYNC=/usr/bin/rsync
# VSERVERINFO=/usr/sbin/vserver-info
# VSERVER=/usr/sbin/vserver
# VROOTDIR=/var/lib/vservers

most of the settings have been left as they are. save the file with a ctrl+k+x in joe.
i’m using rdiff only but there are other options available. see the backupninja wiki for more info.
now to get an action (or backup job) sorted. from the wiki above we’re going to use the action.rdiff file. first we want to cd into the backup.d directory like so. as root ‘cd /etc/backup.d’. then ‘wget http://cloudplasma.co.uk/wp-content/uploads/action.rdiff’ to download the default configuration file. do a ‘joe action.rdiff’ to open it and tweak it to your liking. mine resembles the following…
######################################################
## source section
## (where the files to be backed up are coming from)

[source]

# an optional subdirectory below ‘directory’ (see [dest])
label = box

# only local type is currently supported
type = local

# how many days of data to keep
keep = 185

# files to include in the backup
# (supports globbing with ‘*’)
include = /var/spool/cron/crontabs
include = /var/backups
include = /etc
include = /root
include = /home
include = /usr/local/bin
include = /usr/local/sbin
include = /var/lib/dpkg/status
include = /var/lib/dpkg/status-old

# files to exclude from the backup
# (supports globbing with ‘*’)
#exclude = /home/*/.gnupg

######################################################
## destination section
## (where the files are copied to)

[dest]

# remote or local? If local, you dont need to specify a host below
type = local

# the machine which will receive the backups
host = backup

# put the backups under this directory
directory = /media/backup/ninja

# make the files owned by this user
# note: you must be able to ssh backupuser@backhost
# without specifying a password
# user = col

save it with ctrl+k+x as before. now it should run at 1.00 am every day. you can try it now by running ‘backupnija –n’ as root.
if you run into permission errors make sure that action.rdiff has 400 permissions (‘chmod 400 action.rdiff’) and that joe hasn’t left a temp version of action.rdiff in the backup directory (‘rm action.rdiff~’).

the old machine runs debian 6.0.3 and deals with a few chores…apache, postfix, remote ssh tunnelling, subsonic and a couple of samba shares. the old install is obviously 32 bit…the new one will be 64 bit (also running debian 6.0.3…i haven’t found a more stable server os) so that prevents a straight rsync as many of the packages will be incompatible with the different architecture. a chum got me thinking…a simple way to get a new box running the same apps as the old one is to replicate the packages that are installed. this can be done simply with the following…on the old box issue the following in a terminal…doesn’t need to be as root…

dpkg –get-selections > installedpackages

this creates a file called ‘installedpackages’ that contains all the applications that have been installed on the old system. if you have the new server up and running and accessible via ssh you can then copy the file over with scp…if not use USB

scp installedpackages username@newserver:/home/username

this will plonk it into your home folder on the new machine. so now log into your new machine and as root issue the command

dpkg –set-selections < installedpackages

and then

apt-get dselect-upgrade

this will install all the packages that ran on the old box…to the new one.

NB that this, obviously, doesn’t move your /home directories, email, web sites or any configuration files. you can do this via USB or using the scp method detailed above. lots of the configuration files are stored in /etc, web sites are generally in /var/www. i’m not dealing with things like mysql now but here’s some info on that.

doing the initial update using apt-get update and then apt-get dist-upgrade (as root). a couple of the ‘extra. repositories were unavailable. also neither ai32-libs or ia32-libs-multiarch would install due to dependency problems.

this build of ubuntu is shipping with firefox 9 and thunderbird 9 but by the time april 2012 rolls around both of these will probably have morphed to version 11 with no noticeable improvements.

as i said…they’re sticking with unity. just as feature-rich as always…

not too light on resources either but then i’ve got a bit of ram to spare.

now as i mentioned before…unity isn’t for me. i could have installed gnome 3 but prefer things to be slightly more configurable so here’s how to add xfce as a desktop environment…in other words change ubuntu 12.04 into xubuntu 12.04…

as root issue the command apt-get install xubuntu-desktop this will grab 120MB or so from the repositories…after it has finished log out. when you’re greeted with the log in screen click the wee gear icon next to your username and choose ‘xfce session’ or whatever. this’ll give you something slightly more usable. i’ve used xfce a fair bit in the past…still not 100% sure of it but it kicks unity’s arse…see below for a quick look.

too many icons…not sure why all my drives have to be shown by default. right-click the bottom menu and go into panel options to make it auto-hide.

ah…proper menus…bliss…

and there’s my daily look at el reg…with some thumbnails…

to sum up. ubuntu’s going in the wrong direction for me interface-wise. they won’t change and i don’t think i’ll ever like it. for an alpha version it seems stable though…adding xfce makes it a lot nicer to use. how long will i stick with it? probably until this afternoon.

subsonic should take care of all your media streaming desires. grab it from here. i’ll be using the debian version on my headerless server.

first off install the java runtime environment (shame, that)…as root issue the following

apt-get install openjdk-6-jre

once that’s complete you’ll need to install the binary you downloaded earlier. again, as root…

dpkg -i subsonic-x.x.deb

and that should be it up and running. login to the new installation using your domain or IP at port 4040 (i’m testing from home so for me that’s http://box:4040) and you should be greeted with something like this

username and password are admin (change the password straight after you login for the first time). most important step? add the path to your media…

a directory for your playlists isn’t created during the install procedure so add one now…as root…

mkdir /var/playlists

now to forward the relevant ports on your router so that streaming will work remotely…the default is, as mentioned before, 4040 but you can change that if you like

if you want to stream from afar you’ll need a domain name if you don’t want to use your IP (which will change if you have a dynamic connection). i use dyn.com but it seems their free service doesn’t exist anymore…i could be wrong. you could try no-ip.com.

that’s about it…you can now browse your media, load and save playlists…and play your tunes.

EDIT — i like it. a lot. have donated and i recommend you do too if you find it useful. it’ll give you a constant url for your dynamic IP address…android app that works after the 30 days trial…and help the developer to continue to improve this cracking bit of code.

biggest change? gnome 3. there was much discussion as to whether mint should switch. i’ve tried various implementations of gnome’s new interface and i have to say that mint’s is probably the most usable so far. it retains the lower panel which helps when many windows are open at once. think they’ve made the right decision. caught up with the latest kernel too. see below for some grabs. seems stable…quick enough…and comes with much of the software that isn’t installed by default with ubuntu (gimp, build-essentals and more) although ia32-libs isn’t for some reason. if you plan on playing with 32 bit code then issue an apt-get install ia32-libs ia32-libs-multiarch as root.

fire up the app on your phone and pick a profile. you’ll need to enter your hostname or IP address. the first time i tried to connect it failed…i had to long-press the database entry and choose ‘edit (advanced)’ to add the username and password. in there you can also choose the default directory that will be displayed and choose the interface theme. works well…though i’ve found that no GUI is displayed on my laptop and i couldn’t get vlc to exit from inside the phone app. haven’t fiddled with it too much though. found it a lot easier to use and setup than gmote. some screenshots can be seen below…

apt-get install gnome-shell

when that’s completed log out and then click the small gear icon next to your username and select gnome from the menu. is it better? probably not. not yet.

the install was smooth and rather speedy (it’s still using some of the 11.04 artwork but that’s normal with ubuntu alpha versions) and within 20 minutes i got to the login screen.

ubuntu have decided to stick with unity instead of moving to gnome 3 as fedora have which i can kind of understand as they’ve ploughed so many resources and so much time into it. after logging in we get to the main desktop interface and can load up some ‘goodies’.

i’m not disliking unity as much as i did the last time i used it…the left bar is actually more configurable than gnome 3’s version. apps can be easily removed (here i got rid of all the usual office guff that’s installed by default) and others added from the main application menu which looks like this…

if you’re too lazy to fire up a terminal window the software centre makes it easy to find and install new software (i.e. first stop… gimp, build-essential, joe, ubuntu-restricted-extras, filezilla and ufraw)

11.10 is using a brand-spanking-new kernel (3.0.0–7-generic). hot off the presses. read more about this big step for linux here and here’s the proof…

memory usage isn’t exactly minimalist but i had a few things running and the virtual machine was only allocated 1GB of RAM. here’s top and htop displaying the current load and processes

i’m actually quite impressed with 11.10 so far. it’s rather polished and reacts snappily (even when running virtually) and while i wouldn’t recommend you install it on a mission critical machine right now i’m just about to do just that.

EDIT — install completed on my HP Notebook. ran into one problem after logging in… ‘/usr/lib/libgconf2-4/gconf-sanity-check-2 exits with status 256′ this also happened when i installed the 10.10 beta as i documented here. follow the instructions in that post to correct the error. here’s a wee grab of the finished article running on my machine…

well packaged, eh?

good recycled padding there…

and the seller is a daily mail reader…no accounting for taste i guess.

and there she is. condition’s fine. as i said earlier…just the power supply included…luckily i have many bags of cables lying around. one of which had the appropriate usb lead.

to get to the film scanner one must remove the white background from the lid of the flatbed. this wasn’t instantly obvious to me. little bit of swearing was necessary before i realised this. the scanner head was also a trifle stuck. a good dunt fixed that.

linux drivers. i only bought the epson as i knew they had some linux support. check their site here to locate and download the correct file. a surprisingly wide range of linux distros are supported. i’m using fedora 15 64 bit so went with the rpms (iscan-2.10.0–1.c2.i386.rpm & iscan-plugin-gt-f500-1.0.0–1.c2.i386.rpm) although the source is there if you want it. install them in the usual manner…there’s a pdf there too if you’re struggling.

now i don’t ever buy software but i’d used vuescan on windows years ago and liked it a lot. i figured that i’d got the scanner for a good price and so $39.95 for some cross-platform software wasn’t too steep (you could just use xsane of course). this has a year of updates…$79.95 for a lifetime license.

here’s vuescan doing its stuff. simple and effective. scanned a strip of 35mm ilford hp5 plus first…

and here’s one of the frames…

not that great but that’s the fault of the photographer…not the hardware of software.

so. for £40 or thereabouts i’ve got a very capable film scanner, that’ll do prints nicely also, and some powerful software. not too shabby i think.

cp –r .mozilla .mozilla.bak

download the latest version of firefox aurora from the mozilla site here

cd into wherever you downloaded this file to…for me it’s

cd Downloads

now, extract the file to a directory called .firefox in your home directory

tar –jxvf firefox* –C /home/yourusername/.firefox

next we need to install some libraries that fedora 15 doesn’t install by default. as root issue

yum install libstdc++.so.* glibc-2.14–3.i686 xulrunner-5.0–2.fc15.i686

close down any firefox windows you have open and run the following to start firefox (version 6.0a2 at the time of writing)

.firefox/firefox

there may be easier ways of doing this…i’m not a linux guru. if you have your own tips feel free to comment. some of my extensions aren’t working…most notably xmarks but this will improve in time. seems stable…i’m using it to write this post in fact.