Skip to content

Unix tools

Created: 2016-02-16 16:29:39 -0800 Modified: 2023-12-15 15:46:50 -0800

  • ”usr” stands for “Unix System Resources”, not “user”. 😜

I don’t know where else to put this, but ctrl+S will usually freeze the terminal’s output so that it doesn’t spam so much, then pressing ctrl+C will let you terminate it without having to fight for priority. Afterward, you may need to press ctrl+Q to unfreeze the output. Tenbroya shared this tip with me.

This is useful for seeing CPU/memory usage across processes/threads

  • ”$ htop -d 1” to increase the refresh rate (X is in tenths of a second)
  • Shift+H: collapse processes so that they don’t show separate entries for each thread
  • You can press tab to go to the I/O section.
  • You can see network and disk I/O by pressing F2, going to “Meters”, and adding network and disk I/O to the second column.
  • Alternatively, use btop for some cool-lookin’ charts.

Schedule something to run in a certain amount of time.

  • Run something in 5 minutes:
    • at now +5 minute
      • It’ll then prompt you for some commands. You can type something like “touch ~/five_minutes_done.txt” and then type ctrl+D
  • Alternative way of running something:
    • echo touch ./hi.txt | at 1:30pm
  • Run something in 5 seconds (reference):
    • echo “sleep 5; COMMAND” | at now
  • See jobs: atq
  • Remove job from queue: atrm job_number

This is useful for stress testing. There’s a Docker image here. Installation/usage (I did these steps from Alpine Linux):

wget http://people.seas.harvard.edu/~apw/stress/stress-1.0.4.tar.gz

tar -xvf stress-1.0.4.tar.gz

cd stress-1.0.4/

./configure

make

make install

find . -name “node_modules” -exec rm -rf ’{}’ +

About the curly braces (taken from the man page):

The string {}' is replaced by the current file name being processed everywhere it occurs in the arguments to the command, not just in arguments where it is alone, as in some versions of find. Both of these constructions might need to be escaped (with a ) or quoted to protect them from expansion by the shell.

About the ”+“

[09:20] Bugzorr: The + adds as many files / directories as possible and a semicolon on the end will run the command for each found file / directory separately

[09:25] Bugzorr: For next time; the -delete flag might’ve worked better than the -exec rm command because it implies -depth. That would won’t generate the errors

Didn’t actually look into this yet, so here are some notes from kensodev:

http://www.theunixschool.com/2012/04/what-is-cdpath.html

11:43 kensodev: @Adam13531 look up $CDPATH

11:43 kensodev: when you do cd some-folder

11:43 kensodev: it will always do the right thing

11:43 kensodev: even when you are in that folder already

11:44 kensodev: say you have ~/Code/adam

11:44 kensodev: when you do cd adam

11:44 kensodev: it will look for that in Code

11:44 kensodev: even if you are in a completely different dir

11:44 kensodev: very useful

SSH probably isn’t running on the server:

  • sudo apt-get update
  • sudo apt-get install openssh-server

For the client, to quickly set up an SSH key and put it on the server (i.e. this is just for quick testing):

  • Client
    • ssh-keygen
      • Generate into the current directory
    • Copy contents id_rsa.pub of id_rsa.pub into ~/.ssh/authorized_keys
      • echo ssh-rsa blahblahblahblahblah name@computer >> ~/.ssh/authorized_keys
    • scp -i ./id_rsa <rest of command>

Putting a file on a remote machine

Section titled Putting a file on a remote machine

Basic (no SSH key needed):

scp file.txt user@ip_address:/home/user/destination/

SSH key specified:

scp -i B:CodeJavaScriptlearningawsFirstkeypair .pem a.txt ec2-user@52.36.170.141:/home/ec2-user

Getting a file from a remote machine is basically the same general way

Section titled Getting a file from a remote machine is basically the same general way

scp root@104.236.198.88:/root/BotLand.zip ./

I couldn’t figure out how to get directories with spaces in them, so if that happens to you, just move the target file first to a directory without a space.

Note: actually formatted for AWS:

scp -i B:CodeJavaScriptlearningawsFirstkeypair .pem ec2-user@35.163.71.233:/home/ec2-user/b.txt ./

Caveats:

  • On Windows, you can’t seem to specify paths with “C:” in them, otherwise you’ll get an error that says “ssh: Could not resolve hostname c: Name or service not known”. I fixed this by just always using relative paths. You can further improve this by just using “CD /D <absolute path to folder>” right before using a relative path.
  • You cannot use SCP just to create folders on the target machine like you can with “mkdir -p”, so you should use SSH just before it, e.g.

ssh -i D:CodeJavaScriptlearningawsFirstkeypair .pem ec2-user@52.36.170.141 “mkdir -p hello/world”

Note: ideally you would specify an absolute path.

Make a temporary directory. This is especially useful when saving directly to a variable:

temp_dir=$(mktemp)

chmod 1777 /var/tmp

The “1” will set a bit which only allows the owner of a file to delete it. So when you “ls -al /var/tmp”, you’ll see “drwxrwxrwt” listed.

If you try to use a private key file via “ssh-add” and you see an error that says “WARNING: UNPROTECTED PRIVATE KEY FILE!”, then you should probably do “chmod 0600 <file>“.

Changes owner of a file. This is useful if you had to copy a file from another user’s home folder on your machine.

sudo chown adam <file>

”curl -X POST” will force a post request so that you don’t need to provide an empty ‘data’ argument like this: curl —data "" url.com

”ls -h” will print human-readable file sizes.

Apparently some people will alias “ls” to “ls -al” so that you don’t need to type something like “ls -alh” every time you want to see human-readable sizes.

$ time <some other command>

This will show how long that particular command takes.

E.g. ”$ time sleep 2” will show a “real” time of about 2 seconds.

Example: sudo -iu bldeploy

  • ”-u” specifies the user.
  • ”-i” (lowercase “eye”) specifies that you want to run in an environment similar to if you’d logged in as that user. This will look at /etc/passwd to find out your login shell, then based on the shell will “source” your various profiles, e.g. ~/.bash_profile.
  • This is much easier than doing “sudo su bldeploy”, because then you would need to run “bash —login” afterward.

Use this to figure out what is open on which ports: “netstat -tulnp”. Use it with “sudo” to see the program names.

Continually runs/prints the output of a command:

  • Run “ls” every 3 seconds:
    • watch -n 3 ls

This is installed via apache2-utils (reference). With this tool, you can generate htpasswd files for use in authentication. These files can technically be shared publicly because they’re salted/hashed passwords.

To generate a new file:

htpasswd -n -B -C 20 bldeploy

-n: display results on stdout

-B: use bcrypt (very secure)

-C: set computing time used for the bcrypt algorithm. Higher is slower but more secure.

bldeploy: the username to generate the password for

After doing this, make sure to clear your Bash history!

To verify a password that you’ve set:

  • Save the htpasswd file as something (e.g. just a file named “p”)
  • htpasswd -bv -C 6 ./p bldeploy <password from the generation step>

Basic usage to replace via regex in a file:

sed -i ‘s/original/new/g’ file.txt

Note: on macOS, you need two apostrophes after “-i” like this (reference): sed -i ” ‘s/foo/bar/g’ “test.txt”

Basic regex substitution with a string:

Input: sed -E “s/(fail|bad)/success/g” <<< “bad, such fail”

Output: success, such success

In Bash, you can modify the contents of a file temporarily using a temporary buffer (reference), that way you don’t need to output to a temporary file.

sudo apt install dnsutils

You can use “tee” to both output to a file and STDOUT at the same time.

Sample usage:

echo hello | tee hello.txt

↑ Prints “hello” immediately and creates “hello.txt” with “hello” in it.

Cheat sheet: https://gist.github.com/MohamedAlaa/2961058

Great starter configuration for all sorts of things (keybindings, features, etc.): https://github.com/gpakosz/.tmux

  • Make sure that when doing “ln”, you use the absolute path to .tmux.conf

grep -r -n —color —exclude={bundle*} —exclude-dir={node_modules,.git,dist,logs} “lodash” ./

  • -r: recursive
  • -n: show line number
  • —color: show colors in output
  • —exclude: exclude files
  • —exclude-dir: exclude directories
  • ”search term”: the thing to search for
  • ./: the directory to start the search in
  • -i: case insensitive (off by default, meaning searches are case-sensitive)

I had something like this:

FOO: ‘FOO’ <— I want to find that

FOO: ‘BAR’ <— I don’t want to find that

BAZ: ‘BAZ’ <— I want to find that

This involved using a backreference:

grep -E “(w+): ‘1” ./main.js

The “-E” is to use extended regex

”grep -v” inverts all of the matches. For example, if you do a search and it returns 1 result from DirA and 100 results from DirB and you only care about DirA, you can do

<original grep command> | grep -v DirB

Getting colorful results by default

Section titled Getting colorful results by default

You used to be able to set an environment variable that would add “—color=auto” to every grep command, but they deprecated that. Now, you have to make an alias that has the option specified by default:

alias grep=“/usr/bin/grep —color=auto”