SSH Workflow Tips

Here is a small collections of tips and tricks with the intention to help improve usability and general productivity when ssh-ing into a linux machine. It was written with Ubuntu/Debian in mind, but should also work on other linux versions such as Mint.

What this post covers:

1. Use history and grep
2. Use the command line faster
3. Use nano for config files
4. Some useful applications
5. Customize your .bashrc file

For the most part these are all fairly straightforward and this post will probably contain nothing new for anyone that has a fair amount of linux experience.

1. Use history and grep

You may want to look up some commands you ran a while back (for example if you want to repeat the same steps on another server or VM).

If you type history you will get a list of all commands you’ve typed recenctly, including past sessions.

You can pipe that output into the grep command to filter out all the commands that had a specific string. So for example if you want to find every command you’ve ever run containing “test”, you can do history | grep "test":

Note that each user has their own history, so if you use a command with sudo, such as sudo apt-get install ... the command will only show up in the root user’s history.

Running sudo history may result in a sudo: history: command not found error. To get around this you’ll need to elevate yourself to be the root user first:

sudo su                           # You are now root
history | grep "apt-get install"  # Running history as root

2. Use the command line faster

Some basic, but useful tips to make using the command line far less tedious:

Use tab to autocomplete

When you are navigating a complex folder structur, you should always press tab after typing the start of the path you want to travel.

For example, if you want to navigate to:

~/ReallyReally/VeryLong/AwesomePath

You can do so easily by typing cd ~/R tab V tab A tab:

Assuming that at each level the name of the directory is unique. Each time you hit path, the console would attempt to auto-complete as much of the path as possible.

If you had both ~/ReallyReally and ~/ReallySorta then typing ~/R tab would autocomplete to where the two names differ, at which point typing either R tab or S tab would complete that directory name. When it stops you can hit tab again to see all possible options.

This is different from the behavior of the Windows commandline, which auto-completes the first option, and each tab cycles to the next option.

Use !!, !$ and !!:n to quickly repeat commands

The !! macro is replaced by your last command, allowing you to quickly re-run commands with sudo:

Similarly you can grab only the arguments to the last command with !$, so you could use that to re-run just a command that followed sudo.

Lastly, if you ever want to grab a specific argument from the last command, you can use !!:n to grab the n’th argument:

Use Ctrl+R to quickly run past commands

As in the windows commandline, you can press the up and down buttons to navigate your most recently run commands.

However if you type Ctrl+R you can get an auto-completing reverse-i-search. Simply start typing and the best matching command will fill in, then press enter when you have selected the right command.

Repeatedly hitting Ctrl+R will cycle through all commands that might match, even with fuzzy searching:

3. Use nano for config files

nano is probably the easiest linux text editor to get into, and ideal for small config changes. It’s what I’ll be using in the rest of this tutorial, and there are some small changes you can make to make it more usable.

While learning more full-featured editors such as vi is encouraged if you plan to do serious development on linux servers, nano is really easy to get a hang of.

Basic nano Commands

• Creating/Opening a file: nano <filename>
• The standard UI:
  • The commands are listed at the bottom, and all are of the format Ctrl+Letter (^ stands for Ctrl)
• Ctrl+O to save document.
• Ctrl+X to close document.
  • If pending changes type y and enter, then when prompted for filename hit enter again.
• Ctrl+W to search for a specific string in the document,
• Ctrl+K to ‘cut’ the current selection (default whole line) Use Ctrl+U to ‘uncut’ (copy and paste)
• Ctrl+- or Ctrl+_ to jump the cursor to a given line number.

IMPORTANT: Ctrl+Z will minimize an open document without actually closing it. This is an easy mistake to make when first starting to use nano. To bring it back to the foreground use the command fg

Some nano Configuration changes

Some small changes can be made to the default nano configuration found in /etc/nano/

sudo nano /etc/nanorc

(Remember, you can use Ctrl+W to search for them)

• set constantshow : you always see the current line number in the bottom bar.
• set softwrap : if you want to see long lines in one screen
• set tabsize 8 : uncomment and change to whatever value you prefer.

4. Some useful applications

Use htop to view running processes

htop is a fantastic interactive process and system resource viewer. You can view all currently running processes, as well as system Memory Usage, CPU usage, Load Average and system uptime.

Simply install it with:

apt-get install htop

And run it by typing htop, and close it with F10 or the standard Ctrl+C.

Some useful commands:

• F6 you can choose what column your processes are sorted by, allowing you to easily identify CPU or Memory hogs:
• F5 to toggle between sorted mode and tree mode (where child processes are stacked below parent processes)
• Move the cursor up and down with the arrow keys and hit F9 to send a signal, such as Sigterm to kill a process.
• F4, type in a name to filter and hit enter to only display certain processes
• F2 and use the arrow keys to navigate the configuration settings.

Note: Certain processes will only be visible to root, and can only be killed by root, so when necessary run sudo htop instead.

Use supervisorctl to easily daemonize anything

If you ever want to daemonize a script or application so that it runs without supervision, this can be a pain without specialized applications such as pm2 for node.js applications.

Enter supervisor, a powerful solution for managing and running processes without having to resort to writing startup scripts or using screen.

It is far too large to cover here, but it’s so painless and useful compared to really any alternative that it’s worth mentioning. I personally use it for running several game servers that don’t provide straightforward daemonization.

Here is a good tutorial to get you started if you want to use it: https://www.digitalocean.com/community/tutorials/how-to-install-and-manage-supervisor-on-ubuntu-and-debian-vps

Note: for node.js applications PM2 should be used instead

Use TheF*ck when frustrated

We’ve all been here:

AGH!! Yes just do it as root already, you know what I meant!

While you could use the !! macro to simply re-run the command, you can instead install TheFuck to automatically fix your command, and vent some frustration at the same time:

It will intelligently try to auto-complete the most recent command you typed. It works for a ton of other commands as well.

It is a python script that you can install via pip like this:

sudo apt update
sudo apt install python3-dev python3-pip
sudo -H pip3 install thefuck

And then you will need to edit the .bashrc startup file located in your user’s home directory:

nano ~/.bashrc

Navigate to the bottom and add the following:

# Default TheFuck Alias (type 'fuck' to activate)
eval "$(thefuck --alias)"
# Custom TheFuck Alias (type 'FUCK' to activate)
eval "$(thefuck --alias FUCK)" 

And then re-load the startup file, either by disconnecting and reconnecting, or by running

source ~/.bashrc

5. Customize your .bashrc file

Speaking of .bashrc, it is a startup script located in a user’s home directory that is run when they log in. It sets up settings, such as what the part of their console to the left of the $ looks like.

Any change you make to a .bashrc file will only take effect when it is reloaded, either by relogging or typing source path/to/file.

NOTE: each user has their own .bashrc file. For example root’s is located in /root/.bashrc instead of /home/<username>/.bashrc. It is what is loaded when you switch to root via sudo su

Add custom command aliases

You can define custom command aliases easily with the format:

alias <name>='<command>'

For example, if you are used to using dir in the windows commandline and can’t stop accidentally typing that instead of ls -la, you can add this alias:

alias dir='ls -la'

Instead of typing sudo apt-get update and sudo apt-get upgrade you could define just update and upgrade:

alias update='sudo apt-get update'
alias upgrade='sudo apt-get upgrade'

Change your PATH

If you plan on installing custom commands and want to make sure they appear in your path, .bashrc is where you do so.

Let’s say you want to add /home/username/testing to your path. You would include the following line at the bottom of your .bashrc file (on linux the PATH variable is delimited with colon : characters))

export PATH=$PATH:/home/username/testing

Which would append your path to the end of the PATH variable.

If you want to replace something already on the path, you’ll want to prepend your path as follows:

export PATH=/home/username/testing:$PATH

To make sure that it is executed first before the path you are trying to replace.

Increase your history size

As mentioned earlier, the history command is very useful for looking back at what you’ve previously done on a linux machine. The bashrc file defines two variables that determine how this data is stored:

HISTSIZE=1000      # the number of most recent commands that will be written to the history file
HISTFILESIZE=2000  # the number of total commands the history file will store

While it is probably unlikely for you to run more than 1000 commands, it’s pretty easy to reach more than 2000 commands total if you frequently run commands on your server. Therefore it’s a good idea to increase the HISTFILESIZE to something larger like 50000. Since the only data that is stored is the command, not the resulting console output, the total size of the file will still be really small.

Enable console coloring

Adding some color to your console can greatly improve the experience of working in the console.

newer versions of debian/ubuntu will already add some color to their output, but otherwise this is done by modifying the PS1 variable, which is used to alter the appearance of the console string.

First, you’ll want nano ~/bashrc and change this line:

#force_color_prompt=yes

to:

force_color_prompt=yes

Then the PS1 variable should be set as follows (note that on most recent ubuntu and debian installations this is the default):

if [ "$color_prompt" = yes ]; then
    PS1='${debian_chroot:+($debian_chroot)}\[\033[01;32m\]\u@\h\[\033[00m\]:\[\033[01;34m\]\w\[\033[00m\]\$ '
else
    PS1='${debian_chroot:+($debian_chroot)}\u@\h:\w\$ '
fi

This will change your commandline from this: To This:

Additionally, this can be used to make it easier to tell when you are using root, by showing them as a different color:

To do this, first follow the same steps as above with root’s .bashrc file with sudo nano /root/.bashrc. (remember to also increase its HISTFILESIZE if you haven’t yet).

Then once color is enabled, you can change the color from green to red by changing the first 32m in the PS1 variable to be 31m instead.

So for your root’s .bashrc the line would now be:

if [ "$color_prompt" = yes ]; then
    PS1='${debian_chroot:+($debian_chroot)}\[\033[01;31m\]\u@\h\[\033[00m\]:\[\033[01;34m\]\w\[\033[00m\]\$ '
else

Show git repository branch

When working with git repositories it is really useful to be able to see when you are inside a git repository, and which branch has been checked out in your working copy, like this:

To do this you first need to include this function at the bottom of your .bashrc file, which will be used to grab the current branch string.

# Git branch in terminal prompt.
parse_git_branch() {
    git branch 2> /dev/null | sed -e '/^[^*]/d' -e 's/* \(.*\)/ [\1]/'
}

Then you can modify the PS1 variable to replace \[\033[00m\] with \[\033[00m\]\[\033[32m\]$(parse_git_branch)\[\033[00m\]

For non-root it should then look like:

if [ "$color_prompt" = yes ]; then
    PS1='${debian_chroot:+($debian_chroot)}\[\033[01;32m\]\u@\h\[\033[00m\]:\[\033[01;34m\]\w\[\033[00m\]\[\033[32m\]$(parse_git_branch)\[\033[00m\]\$ '
else

Note: You’ll once again want to apply that change to root’s .bashrc file as well, and make sure to change that first 32m to 31m to ensure its color remains red..