Databases, Systems & Networks

Source: Juan Valencia’s website

In this post I cover the basics of rsync, in preparation for a subsequent post that will cover backups and it’s use in conjunction with cronjobs to automatize the backup process. From the copying and synchronization of local files and folders, to it’s use for transfer information among computers. Itsuse as a daemon when SSH is unavailable was moved to it’s own section.

Topics
The basics of rsync
Copying local files and folders
Dealing with whitespace and rare characters
Update the contents of a folder
Synchronizing two folders with rsync
Compressing the files while transferring them
Transferring files between two remote systems
Excluding files and directories
Running rsync as a daemon (moved to it’s own section)
Some additional rsync parameters
Footnotes

The basics of rsync

rsync is a very versatile copying and backup tool that is included by default in almost every Linux distribution. It can be used as an advanced copying tool, allowing us to copy files both locally and remotely. It can also be used as a backup tool. It supports the creation of incremental backups.

rsync counts with a famous delta-transfer algorithm that allows us to transfer new files as well as recent changes to existent files, while ignoring unchanged files. Additionally to this, the behavior ofrsync can be throughly customized, helping us to automatize backups, it can also be run as a daemon to turn the computer into a host and allow rsync clients connect to it.

Besides the copying of local files and folders, rsync allow us to copy over SSH (Secure Shell), RSH (Remote Shell) and it can be run as a daemon in a computer and allow other computers to connect to it, when rsync is run as a daemon it listens to the port TCP 873.

When we use rsync as a daemon or when we use RSH, the data that is send between computers travels unencrypted, so, if you are transferring files between two computers in the same local network, this is useful, but this shouldn’t be used to transfer files over insecure networks, such as the Internet. For this purpose SSH is the way to go.

This is the main reason why I favor the use of SSH for my transfers, besides, since SSH is secure, many servers have the SSH daemon available. But the use of rsync as a daemon for transfers over fast connections, as is usually the case in a local network, is useful. I don’t have the RSH daemon running in my computers so you may find me a bit biased about SSH in the examples. The examples covering the transfer of files between two computers use SSH as the medium of transport, but in a separate post I cover the use of rsync as a daemon.

Lire la suite…

It can be helpful to configure Mac OS X to automatically mount shared network drives, this is particularly true for those of us who regularly connect to a network drive for file sharing or backups.

Setting up automatic network drive connections in OS X is a two-step process, you must mount the drive, then you add it to your automatic login items. This should work flawlessly in most versions of OS X, but we’ll cover an alternative approach that uses Automator to mount a network drive automatically on login as well.

1) Mounting the Network Drive

If you’re already familiar with mapping a network drive in Mac OS X you can skip the first part of this and go straight to System Preferences in the second section.

1. From the OS X desktop, pull down the “Go” menu and select “Connect to Server”
2. Connect to the server and mount the drive you want to automatically connect to on boot
3. Choose Guest or for a specific user check the box next to “Remember this password in my keychain” – you must select to remember the password otherwise the automatic login event can not happen without logging into the network drive

Next, you add the network drive to automatically connect on OS X by bringing it into your Login Items list.

Lire la suite…

Copy files or directories from one location to an another host by rsync.

If you’d like to set rsync automatically by cron or others, it need to configure like follows because authentication is required without settings. For example, Copy files or directories under the [/root/work] on dlp.srv.world to [/home/backup] on www.srv.world.

[1] Configure on source host.

root@dlp:~# apt-get -y install rsync
root@dlp:~# vi /etc/rsync_exclude.lst
# specify files or directories you'd like to exclude to copy
test
test.txt

[2] Configure on destination host.

root@www:~# apt-get -y install rsync
root@www:~# vi /etc/default/rsync
# line 8: change
RSYNC_ENABLE=true
root@www:~# vi /etc/rsyncd.conf
# create new
# any name you like
[backup]
# destination directory to copy
path = /home/backup
# hosts you allow to access
hosts allow = 10.0.0.30
hosts deny = *
list = true
uid = root
gid = root
read only = false
root@www:~# mkdir /home/backup
root@www:~# systemctl start rsync

[3] It’s OK. Execute rsync on Source Host like follows.

root@dlp:~# rsync -avz --delete --exclude-from=/etc/rsync_exclude.lst /root/work/ www.srv.world::backup
# Add in cron if you'd like to run reguraly
root@dlp:~# crontab -e
# for example, run at 2:00 AM in a day
00 02 * * * rsync -avz --delete --exclude-from=/etc/rsync_exclude.lst /root/work/ www.srv.world::backup

Sauvegarde MySQL

Pour sauvegarder une base de données (sans et avec compression) :

# mysqldump NOM_BASE > NOM_FICHIER
# mysqldump NOM_BASE | gzip > NOM_FICHIER

Pour restaurer une base de données (sans et avec compression) :

# mysqladmin create NOM_BASE
# mysql NOM_BASE < NOM_FICHIER
# gunzip < NOM_FICHIER | mysql NOM_BASE

Sauvegarder toutes les bases :

# mysqldump --opt --all-databases > NOM_FICHIER

Pour sauvegarder uniquement certaines tables :

# mysqldump NOM_BASE NOM_TABLE0 [NOM_TABLE1...] > NOM_FICHIER

Pour presque faire un « –exclude » (qui manque cruellement à mysqldump):

mysql -B -N -e 'show databases' | 
  perl -ne 'print unless /b(?:phpmyadmin|mysql|information_schema)b/' | 
  xargs echo mysqldump -B

Et pour sauvegarder des tables correspondant à un motif (préfixe le plus souvent) :

# mysqldump NOM_BASE $(mysql NOM_BASE -B --column-names=False -e "show tables like 'exemple_%'") > NOM_FICHIER

Pour dumper avec une condition particulière :

mysqldump -t <base> <table> --where="my_id='66666666'"

Ce qui permet de réinjecter des données résultantes d’un SELECT * FROM base.table WHERE my_id='66666666'.

Il est évidement possible de faire toutes ces opération sur une instance en précisant son port avec l’option –port (valable pour mysqldump et mysql).

Pour obtenir une liste des utilisateurs mysql, on peut utiliser cette fonction (glanée sur serverfault) :

mygrants()
{
  mysql -B -N -e "SELECT DISTINCT CONCAT(
    'SHOW GRANTS FOR ''', user, '''@''', host, ''';'
    ) AS query FROM mysql.user" | 
  mysql | 
  sed 's/(GRANT .*)/1;/;s/^(Grants for .*)/## 1 ##/;/##/{x;p;x;}'
}

Lire la suite…

1. Introduction

rsnapshot is a backup tool written in Perl that utilizes rsync as its back-end. rsnapshot allows users to create customized incremental backup solutions. This article will discuss the following: the benefits of an incremental backup solution, rsnapshot’s installation, its configuration, and usage examples.

2. Back-it up!

I was recently discussing with a colleague the benefits of backing up your data. My colleague was telling me how one of her customers had recently lost a rather lengthy article that they had been working on. I decided that this may be a good chance to experiment with my netbook and rsnapshot. For this tutorial, I’ll assume you have have 2 pieces of hardware: your host computer, and your destination equipment. I’ll be using an external hard drive for the majority of this post. However, I will briefly cover usage for backing up files over a LAN.

Backing up your data should not be the question to ask but rather how should I backup my stuff? What’s the best way? Well there are many different backup pathways you can take, including block level (dd, partimage), partition level (RAID and all its variations), file level (rsyncand its children applications). I’ll discuss two types of backups in the context of file-based backups.

Normal backups, or full backups, are self explanatory. Normal backups are one way of backing up ALL your files every time you perform a backup. One issue with utilizing a multiple normal backup scheme is that a normal backup takes up a considerable amount of space. For example, if you perform a full backup of a 250gig hard drive at 20% capacity, everyday for just one week (assuming that the amount of data does not fluctuate) will mean that you already have used 350gigs for only one week’s worth of backups. As you can see, that is not feasible in the long run. The other method that I prefer is the incremental backup method. An incremental backup consists of one full backup and then performing additional backups. These additional backups will only backup files that  have changed since the last backup. Instead of backing up your entire hard drive, only the specific files that have changed since the last backup are backed up. As you can probably imagine this is a much more efficient process. One tool that does this on *nix is rsnapshot.

3. What is rsnapshot?

rsnapshot, as mentioned earlier, is an incremental backup utility. In this tutorial, I will show you how to create a seven day rotation backup scheme using rsnapshot. Essentially, rsnapshot will create 1 full backup and then subsequent backups will backup only files that have changed. The true power of rsnapshot is its ability to utilize hard-links between each backup. Each backup will appear to be a full backup. In reality each new backup consists of newly created or updated files. rsnapshot can be used over a LAN and can also be ran from cron. In this tutorial, I’ll show both usage examples.

4. Installation

Installation of rsnapshot is pretty straightforward, simply run the following command:

On Debian (or Ubuntu):

apt-get install rsnapshot

On Fedora:

yum install rsnapshot

On ArchLinux:

pacman -S rsnapshot

Now let’s configure rsnapshot.

Lire la suite…