If you didn't catch the previous article, click here
We will now talk about some security principles & guidelines. We won't be getting very specific in this post, but this will help you get a few things inside your head.
These can be used when you don't know what to do -security wise-. Ask yourself questions from the principles/guidelines below, and do what you think best! Do not be afraid to search around the web and ask!
I've already mentioned some of them in the introduction post, but it will be great to have almost everything gathered here!
If you don't need a piece of software, do not install it. If you installed it, used it once but you won't use it again, remove it.
Some packages includes what we refer to as a "service" or a "daemon". A process that runs in the background either waiting for input (webservers & mailservers would be a great example), or does things when it needs to (cron jobs).
The more services you run and packages you install, the more security wholes you may be introducing!
If you are running various services that are critical for you, make sure to run them in different systems. Either use different servers, of learn about virtualisation and put each service in a different container. The most popular virtualisation systems are OpenVZ, KVM, XEN. Do some research and find which one is better suited for your case. If for example you run a web server with your blog and a VPN to securely access some other parts of your server, if you web server gets compromised, chances are that the attacker will access your VPN as well.
Using secure protocols to do your job (HTTPS, RSYNC, SSH, SFTP) will minimise the chances that you will get compromised. On the other hand, using non secure protocols (such as HTTP, TELNET, FTP) leaves you exposed to eavesdropping and man-in-the-middle attacks.
If there are multiple people accessing a server, each one should have their own account. The same goes for services. Make the extra mile to give Apache2 its own user instead of running it on the same user account with any other software.
Also, accountability is thrown out of the window. As an administrator, if you try to hold accountable user A because he overlooked a detail, and the account is not shared, he will not do it again. But if the account is shared, he will think twice before overlooking something.
Do not directly use the root account! This is a no-brainer. Create a standard user account and give root access to yourself by some other way (note: if you create an administrator account for you, it is the same thing as using the root account)
ALWAYS login as a standard user.
Also, be sure to not give root access to any daemons or people who may not know what they are doing.
This is a usually-overlooked section.
Disable remote logins for service accounts. You don't have to login as "www-data" on your server. You can login with your account and then switch to "www-data".
Disable and/or remove accounts that people no longer use. It will take you a few secods to recreate or re-enable that account if it's needed again.
Using multifactor auth can greatly increase your server/computer security.
Add OTP (one time passwords), that are either generated and sent via Email or SMS to the user, or from an app such as Google Authenticator or Authy.
Enforce strong passwords! Using the password "IamDimitrisAndIloveSteemit" may seem secure, but there are various brute-forcing methods that a hacker could generate that password immediatelly. Instead, substituting letters for numbers and adding special characters can and will make a difference. For example "1amD1mi7ri54nd1l0v3St33m17$" is not so easily guessed. Also, using longer passwords also adds to security.
Another way would be to disable password logins for SSH, and add key-based auth.
The principle means giving a user account only those privileges which are essential to perform its intended function. For example, a user account for the sole purpose of creating backups does not need to install software: hence, it has rights only to run backup and backup-related applications. Any other privileges, such as installing new software, are blocked. The principle applies also to a personal computer user who usually does work in a normal user account, and opens a privileged, password protected account (that is, a superuser) only when the situation absolutely demands it.
Some examples:
Give the bare-minimum privileges needed for someone to do their job.
Escalate to root with sudo when needed, instead of logging as root.
Avoid running services as root user.
Review your log files regularly. If you find nothing wrong, clean them up (logrotate would do the job)
If possible, keep a copy of the old logs to a remote location.
Using a firewall to block incoming connections to services that you don't want exposed is a great way to increase security.
Linux already comes with a powerfull firewall, iptables. It's not very easy to master, but it can do wonders! You can always use a frontend, such as ufw.
Allow incoming connections only from trusted sources. If you have static IP at home, you can restrict connections to your server from that IP address only.
Encryption protects your data when not processed. Linux distros allow you to encrypt your home directory, and they can only be accessed with your key. Take advantage of that!
Thanks for reading Security 101: Principles & Guidelines. Upvote and resteem if you liked my post.
Craving for more? Until the next part is available, have a look at my Server 101 series:
