mjhjmtu

"Can someone tell me what sudo means?" I am learning Linux. I googled but can someone kind of give me a Windows comparison I think this will help me a little. I know su means switch user. sudo is being another user, but how and why? Why can't you just give yourself permissions under your user name?

sudo allows you run a program as the root super user. Using it instead of su,

1. you do not need to know the root password because it asks for your password,


2. you know explicitly what commands will run as the root and which others as you. If you use su, you got a shell and can run inadvertently some (dangerous) programs, and


3. all your actions are logged into the auth.log file. Actions in a shell, such as one initiated by su are not usually logged.

Why can't you just give yourself permissions under your user name?

For security reasons, the root account is seldom used for day to day purposes. Usually, all the users, including administrators, use a "normal" user and only execute commands as root when it is needed. This may prevent (1) unauthorized access to files of other users, (2) not-logged root actions and (3) problems caused by executing "dangerous" programs.

There is an analogy in Wikipedia you can consider:

• Using a root account is rather like being Superman; an administrator's regular user is more like Clark Kent. Clark Kent becomes Superman for only as long as necessary, in order to save people. He then reverts to his "disguise". Root access should be used in the same fashion. The Clark Kent disguise doesn't really restrict him though, as he is still able to use his super powers. This is analogous to using the sudo program.

In early versions of Unix the most common scenario was to be logged in through a terminal without a windowing system. The su command (substitute user) was a convenient way to change your current permissions without logging out. Remember, without a windowing system logging out meant stopping whatever you had running.

Using su to become root has some annoying issues. People occasionally would forget to exit the session and accidentally run commands as root. Changing the administrative password was complicated by the fact that you needed to distribute it to everyone who needed it. Most systems didn't log every command so it was difficult to tell who'd done what as root. And so on.

sudo addressed those issues by letting people run one single command as root. It provides better auditing and control while being a bit more convenient for one off commands.

ok, so let me see if I understand this. Instead of people having admin accounts everyone has the potential to be an admin by using the sudo username. However, by using that sudo user name your actions will be logged back to you because its kind like a hell off your actual username acct. If this is so, I really like that. It saves tons of time. You don't have to wait for someone with privies to come over to you to add software and printers and such.

By default on most Unices, the "administrator" account, root, is to be used only sparingly. There are only two levels of administrator, or "super user", rights on a normal Unix system: You either have full control of everything, or you don't.

"Full control" includes circumventing most permission checks and to utterly mess up the system by a simple mis-typing. As an ordinary user, you can at most delete your own files only.

The older su utility is used together with the root password to assume the root identity, but may also be used to assume other users identities, or to execute single commands as them, given that you belong to the correct user group and know their passwords.

su is commonly used by the root user (on some Unix system) to start daemons (service processes) that need to run under particular user identities to protect them from accessing areas that they shouldn't have access to (typically web servers, SSH server and the like). This may be done in startup scripts as the system is booting.

su is more seldom used interactively (on the command line).

The sudo utility requires a bit more configuration as it allows a user to execute commands as another user by using their own password (which means that a root password, for example, does not need to be shared), and it also allows for fine grained control of who is allowed to do exactly what. One could, for example give a single user, or a whole group of users, the ability to mount a certain disk, but not to unmount it or to mount any other disks, or to install software using specific commands etc. Again, the sudo utility allows for assuming other user's identities (not just root) for interactive sessions, or for executing single commands (if the configuration has been set up this way by some other person with root permissions).

Typically, one would run single commands using sudo (hence the name "superuser do"):

$ sudo apt-get install vim

or

$ sudo shutdown -ph now

Using sudo in this way is prefered to using sudo -s and work with an interactive root shell, because the commands that you execute are logged. one may later retrace the actions in the log file to see where (and who) did something, either for debugging or for auditing.

If you find yourself typing away at an interactive root prompt for an extended period of time, then you should reconsider what you're doing and if that really requires root privileges.

As an example, the only time that I bring up an interactive root prompt is when adding an ordinary user for myself on a freshly installed machine, or when I really have some serious issues that prevents me from even logging in as an ordinary user.

OpenBSD (since release 5.8) has a utility called doas which replaces sudo in the default system installation. It was developed when it became clear that sudo had become too complex to fit the goals laid out by the developers of OpenBSD. Its use is similar to that of sudo, but configuring it is easier.

sudo allows you to run a program as a different user (including root) based on a configurable, predefined set of rules.

On a Unix/POSIX/Linux system root is god. The root user is also singular.

In many environments there are multiple admins who might require full root access, but there can be also be individuals who require access to do some specific function usually reserved for root; DBAs need to stop and start DB services, edit DB config fules, network administrators need to configure network devices and routes, security administrators need to manage certificates and configure firewalls.

By constraining what commands a person can run, sudo applies the principle of least privilege - users don't get access they don't need. But its not just constraining the binaries users run. Many admin tasks are carried out from the command line - and the paramters passed to that command can also be constrained by sudo (or captured in a shell script with access provided by sudo) hence it is possible to restrict the data / devices that these specialized admins can access.

Why can't you just give yourself permissions under your user name?

1) this would be bad.

• I have often had access to the root account on systems - but I aways set up a named account to use for login, and only 'su' or 'sudo' when I want to explitily invoke my super powers. In addition to avoiding silly mistakes, my named account provides a sandbox for any misbehaving components.




• what if there are 5 people who need the privileged access, do I grant it to each and every one? What about 20? what about 200?

2) this would be difficult

• it's not the Unix security model. Sure you could change the permissions on some executables (see earlier note about parameters). But if you're security conscious, you apply patches regularly - which can undermine a customized permissions model (sudo OTOH maintains its permissions independently of the commands themselves).