In this exercise, we will explore the uses and limitations of various tools for creating backup copies of filesystems. Most tools have tasks they are well suited to, and other tasks they get used for but are not well suited to perform. We will create a data directory to use for evaluating tools, then experiment in detail with the simple file copy command to see what kinds of assumptions and limitations must be accounted for. Then we will use rsync to perform backups that have the features we look for in a live backup system.
Begin by running some commands to create a sample data space for us to work in. It will create files owned by multiple users sharing a data space by using group permissions. There are a few directories and different file types in it to simulate actual files you might find on a production server. Simply copy and paste these commands into a shell where you are logged into your Linux VM as student. The last 2 commands will show you what was made. The tree command will show what was made and purports to tell you how much disk space is used by your files. The second runs the du command to tell you how much space is actually used.
mkdir -p ~/data/{a,b,c,d}
pushd ~/data
dd if=/dev/zero bs=100M count=1|base64 >b/bigfile.txt
dd if=/dev/zero of=d/biggerfile bs=1 count=0 seek=512M
for i in {1..3}; do
ln b/bigfile.txt a/altfile$i.txt
ln -s ../b/bigfile.txt c/lfile$i.txt
mknod d/sock$i p
done
sudo chgrp -R staff .
sudo chown -R bob d
sudo chown -R alice a
sudo chown -R ted c
sudo chmod 2750 ?
popd
sudo tree -push --du ~/data
sudo du -sh ~/data
Why is there a difference in the amount of disk space in use between the tree command and the du command? Which one is correct?
In this section we will examine the standard copy program, cp. If you are already familiar with the cp commamd, you may already know what it does. This activity tries using it in multiple ways to uncover limitations in the command that affect the usefulness of files backed up this way. We look at this one because it is commonly misused.
cp without options to see what happens. You should get an error about copying directories.
cp ~/data ~/data2
cp with the recursive option. You should get an error about not having permissions to access files from the student account. Check to see what did get copied, and what that looks like. - screenshot
cp -r ~/data ~/data2
tree -push --du ~/data2
du -sh ~/data ~/data2
cp recursively again, but this time use sudo to solve the permissions problem. Check to see what got created. - screenshot
sudo cp -r ~/data ~/data3
sudo tree -push --du ~/data3
sudo du -sh ~/data ~/data3
cp with the archive option and sudo. Check to see what got created. - screenshot
sudo cp -a ~/data ~/data4
sudo tree -push --du ~/data4
sudo du -sh ~/data ~/data4
cp and other tools like it are designed to copy files from one place to another on a single computer. Backups written to remote computers require additional thought. It is particularly valuable to have versioned backups in order to avoid losing your backups to ransomware. Offline backups can also be useful if done correctly.
In this example, to keep this lab simple, we will just use rsync with ssh to create versioned backups on a remote backup server from our existing Linux VM using the root account. To use this for protection against ransomware or other system compromises, additional steps would have to be taken and we would not use root access. So the main purpose of this lab activity is to become familiar with remote backups created by rsync.
sudo bash
cd
ssh-keygen
cat ~/.ssh/id_rsa.pub
exit
sudo vi /root/.ssh/authorized_keys
sudo ssh root@backup-vm-ip id
date=$(date '+%F-%H-%M') ; sudo rsync -ah --link-dest=/backups/latest --delete --exclude={/proc/*,/tmp/*,/run/*,/dev/*,/sys/*,/mnt/*,/lost+found,/media/*,/backups,/swap.img} / root@backup-srvr-ip:/backups/$date && sudo ssh root@backup-srvr-ip ln -nsf /backups/$date /backups/latest
date=$(date '+%F-%H-%M') ; sudo rsync -ahv --link-dest=/backups/latest --delete --exclude={/proc/*,/tmp/*,/run/*,/dev/*,/sys/*,/mnt/*,/lost+found,/media/*,/backups} / root@backup-srvr-ip:/backups/$date && sudo ssh root@backup-srvr-ip ln -nsf /backups/$date /backups/latest
ls /backups
df -h
du -sh /backups/*/
To restore a backup, we more less reverse the process and copy files from the backups repository to wherever we want to have them. We can use the rsync program or we can just find and manually copy whatever files we want. Perform the following tasks on the original Linux server VM, not the backup server.
mkdir ~/restoredfiles
cd
sudo rsync -av root@backup-srvr-ip:/backups/version-identifier/home/student/ restoredfiles/
Submit one PDF containing screenshots showing the work you did. Everywhere there is a screenshot marker in the instructions, you must capture enough to show the command you ran and the results of running it. Include the question from the section on the cp command with your response.