Common server administration tools

Here is a quick cheat sheet on the various tools we use on the servers. These tools are standard linux tools.

htop

htop is a process watcher. Run it with the simple htop command to see the processes running, and how much CPU and RAM they use. Useful to find processes that are experiencing a high resource use. It is typically the first tool to run after logging on to a server, just to have a general quick look.

podman

Many of the servers run their services in containers, using podman. Note: Containers run under a specific user, make sure to su to the appropriate user for the server you are on, if podman seems “empty”!

Common commands include podman ps -a to list all containers, or podman image ... to handle images, podman pull IMAGE to pull an image from a repository, and podman restart CONTAINER to restart a container.

NOTE: Some servers (gtweb and gtoahpa) have a custom service script to do common tasks, including restarting services running in containers, but using podman (and systemctl) directly is fine, too.

systemd

The main tool to start, stop, restart services. The systemctl command is used to start, stop, restart services, for example sudo systemctl restart <service>. Note the sudo, most systemd-commands must be run with sudo.

less, head, tail, and cat

Remember that some files, particularly log files, can be very large. Loading them up in vim (or even nano) to take a look at, can be very slow. Prefer to use head to look at the start of the file, tail to look towards the end, and less if you want paged output where you can scroll. cat prints the entire file to standard out.

adduser and usermod

Add users with adduser, and add users to the sudo group with usermod -aG sudo USERNAME.

chmod and chown

If permissions are an issue, they are changed with chmod, and chown. Some files need to be executable, do that with chmod +x FILE. Sometimes, a particular user must own the file, do that with chown.

df and du

Use df to see disk usage, and du when looking at file sizes, which includes the sizes of all subdirectories.

rg (or grep)

To search for text in files, all our servers have ripgrep, with the command name rg installed. It is faster than, and has a simpler syntax to remember than grep , which of course also is available.

fd (or find)

To find files by name, use fd, which has a nice and easy syntax, or you can use find, which is also available.

dig

dig is a DNS info query tool. Use it to figure out which domain names are pointing where.

tmux (or screen)

If you are planning to do run a long-running process on the servers, consider using tmux (or screen if tmux is not installed). With those, you can start the command, and then detach (or even log out of the server), and still still have the command running, and attach back to the same terminal session later. Can be useful, so that a command doesn’t stop half way just because the connection to the server failed, for example.

Others

traceroute and ping to check connectivity. netstat, ss, lsof to check check open ports, etc. scp (and rsync can be used to copy files from local to the server). journalctl to look at system logs.