Linux Tutorial

From CAC Documentation wiki
Revision as of 18:49, 6 April 2018 by Pzv2 (talk | contribs)
Jump to navigation Jump to search

This tutorial is intended as a basic introduction to Linux for users of Red Cloud services who are using a Linux Instance. There are two Linux distributions (AKA distros) available for images on Red Cloud: Ubuntu and CentOS. In this tutorial, you will learn how to add a user, install software using the distribution's package manager, how to enable remote password logins, and several related tips. There are some common commands between both distributions, and a section for the specific commands on Ubuntu and CentOS. While many commands are similar across Linux systems, package management, service control, and to a lesser extent, user management, are some of the areas that will differ from distribution to distribution. For a more detailed tutorial, please see the Cornell Virtual Workshop.

Definitions

This section contains some basic working definitions to help you through this tutorial if you have never used Linux before. This list and the definitions in it should not be considered authoritative.

directory - folder

terminal (A.K.A. console or shell) - a text-only user interface for interacting with an operating system's programs and services. This is where commands are entered.

command - a task for the computer to execute that is entered via the terminal

package - an archive of software and metadata that can be downloaded, installed, and removed via a package manager

root - the system administrative account with all the highest privileges, also known as the superuser. By default, most Linux distros have a single root account when installed, and no user accounts.

sudo - a program that allows a user to run commands with the privileges of another user, typically the root account. This is typically used by typing sudo before a command.

Basic Useful Commands

pwd
print working directory
ls
list directory contents
cd
change directory

Example: cd ~ will take you to your home directory

mkdir <name>
make a directory with the specified name
man <command>
manual pages for the specified command
history
displays a list of commands that have been executed via the terminal
cat <file>
outputs the contents of a file to the terminal, with many other options available (check out man cat for more info)
grep <pattern>
prints lines matching a specified pattern. This is usually used with the | command (pronounced "pipe") so that you can "pipe" the output from one command into grep to effectively search it.

Example: history | grep mkdir would search the history output for each time the mkdir command was executed, thus determining all the directories you had created.

Text Editors

Since the default interaction with a Linux Instance is through a terminal, it may be useful to familiarize yourself with at least one text editor that can be used in the terminal. Here are a few, with links to get more information about them, but there are more.

vim
Vim is often already installed with many Linux distros, and is therefore useful to learn. There are many online tutorials, but you can also simply type vimtutor in the terminal to learn how to use vim.
emacs
Emacs is a family of text editors including the very popular GNU Emacs. If you want to use it, it may be helpful to take a guided tour or to consult the manual.
nano
GNU nano is a simpler text editor than something like vim because it doesn't have modes, you simply type when it opens. If you'd like more information, consult the documentation.

Ubuntu

The "ubuntu" user

explain Ubuntu root user weirdness


CentOS

Initial User Setup

Once you have started a Linux Instance, you will want to connect using ssh and create a user account. You will first have to login as the root account and setup the user account yourself. It is advisable to setup the user account instead of continuing to use the root account. This section details how to correctly setup the user account on a CentOS image. For all

  • ssh -i <keyname>.pem root@<ip of instance>
    • Connects to the instance via ssh
  • adduser <username>
    • Adds a new user with the name <username>
    • Note that <username> could be e.g. ‘bob’, it doesn’t need to be (and really should not be) a Cornell netid, since you can optionally configure your instances to allow use of netid and netid passwords for project members
    • Multiple users may be added at the instance owner’s discretion.
  • passwd <username>
    • This will prompt you to set and verify a password for the user
    • Note: if you do not run this command, a password will not be set for the user!
  • mkdir ~<username>/.ssh
    • Creates a directory for the user to hold the public encryption key used in ssh
    • Note: The .ssh folder is hidden to the ls command by default because of the "." at the beginning. You can see all folders by sending the ls -a command.


SSH Security

Once you have set up a user with sudo privileges and ensured that you can indeed login and perform sudo commands successfully (it would be good to test this to be sure), you may want to secure the root login by disabling it.

Disable root login: This must be done while logged in either as root or your user with sudo privileges.

  1. vim /etc/ssh/sshd_config
  2. Change the the line PermitRootLogin yes to PermitRootLogin no
  3. Note: if this line is commented out (with a # character in the front), you will need to uncomment it.
  4. service sshd restart

When you exit, you should verify that you cannot login as root, but that you can still login as your user. For more information, see the CentOS guide on Securing OpenSSH.

Old Tutorial

In this tutorial, you will learn how to add a user, install software from the Linux distribution's package repository, how to enable remote password logins, and several related tips. This tutorial was designed with Ubuntu/Debian systems in mind, but most of the listed commands have the associated Red Hat/CentOS command in parentheses where there is a major difference. While many commands are similar across Linux systems, package management, service control, and to a lesser extent, user management, are some of the areas that will differ from distribution to distribution.

After starting an instance, you will be logged in as the administrative root user. You can create a regular user account:

 adduser <username>
 (redhat: also do passwd <username> to set password)

Note that <username> could be e.g. ‘bob’, it doesn’t need to be (and really should not be) a Cornell netid, since you can optionally configure your instances to allow use of netid and netid passwords for project members. Multiple users may be added at the instance owner’s discretion.

 adduser <username> sudo

This will add <username> to the sudo group, which will enable <username> to easily install software and perform other administrative tasks without needing a root login. This has the advantage of making it more difficult to accidentally do something unfortunate to the system. On Red Hat/CentOS, the process is different; run visudo as root and add the line: <username> ALL=(ALL) ALL

To log in as the new user:
 ssh <username>@localhost


You can replace localhost with the public IP of the machine if you are logging in remotely, instead of from an existing root login on the instance. Controlling SSH authentication, needed for logging in, is beyond the scope of this article, but we will say that for password authentication to work, you may need to change the PasswordAuthentication to 'yes' in /etc/ssh/sshd_config, and restart the ssh daemon. Restarting daemons varies from system to system, but can be done by stopping and starting the instance if it is based on an EBS image.

Alternatively, the recommended approach is to use a public-private keypair. First, as your user, generate a keypair: ssh-keygen -f username creates two files in the current directory; username.pub is the public key, which you could append to the users ~/.ssh/authorized_keys on your server. The username file should be copied (using your root credentials) and sent securely to the user who needs it. That user can then login like so (note the first <username> corresponds to the private key file we generated):

ssh -i /path/to/<username> <username>@localhost

On the server, make sure the user's .ssh directory has correct permissions:

chmod 700 ~username/.ssh


Now we can install an application, but first let’s make sure our package database has all the latest package information by running sudo apt-get update. After this completes, install the screen-saving program tmux, which is often useful in case your connection is dropped (either intentionally or unintentionally) or if you want to have multiple terminals available without needing to login each time:

 sudo apt-get install tmux
  (redhat: yum install tmux)

For information on using tmux, please find a tutorial elsewhere.

To find available packages (available from currently installed repositories), the following command may be used: aptitude search <partial name> (redhat: yum search <partial name>), where is part of the package name. For instance, here are the first 6 results for aptitude search python:

 p   bpython                         - fancy interface to the Python interpreter
 p   bpython-gtk                     - fancy interface to the Python interpreter
 p   bpython-urwid                   - fancy interface to the Python interpreter
 p   bpython3                        - fancy interface to the Python3 interpreter
 p   cairo-dock-plug-ins-dbus-interf - Python interface to interact with Cairo-Do
 p   cantor-backend-python           - Python backend for Cantor

Note that the ‘p’ in the first column means that no trace of package exists on the system (run man aptitude for more details).