usrme/ONELINERS.md

## ONELINERS.md

      
    Raw
  

              ONELINERS.md
            
          
    Oneliners

Bash


Basic for loop to iterate over lines in a file:

for pkg in $(cat pkgs.txt); do sudo apt purge "$pkg" -y; done

More complex for loop using if statement to control actions, clean up output etc:

for node in $(cat nodes.txt); do echo "Node: ${node}"; ssh -q -t "$node" 'if [[ $(lsblk | grep -i lvm) ]]; then sudo apt install mdadm -y; fi'; done

Checking very busy log files for their contents without having your console hang (as opposed to using tail -f):

watch -n 0.5 sudo tail /var/log/named/queries.log

Alternative conditional logic in for loop iterating over array variable:

# Declare 'nodes' variable separately or prepend to loop and separate with semicolon
for node in "${nodes[@]}"; do ping -c 2 -W 0.1 "$node" > /dev/null && echo "OK: ${node}" || echo "NOT OK: ${node}"; done

Use while loop to iterate over lines in a file:

Avoids calls to cat as is the case with the for loop example


# Using 'madison' command rather than 'policy' seems to be slightly faster
while read pkg; do if [[ $(apt-cache madison "$pkg") ]]; then echo "OK: ${pkg} exists in some repo"; else echo "NOT OK: ${pkg} doesn't exist in any repo"; fi; done < pkgs.txt

Match lines into an array:

types=($(grep -oE 'pattern' input.txt))

Grab block of text between two patterns:

sed -n '/pattern1/,/pattern2/p' input.txt

Just see octal permissions for a file or directory:

stat -c '%a' /etc/passwd

Grab last character from string:

last_character=${string_variable:-1}

Parse file list from output of grep into subsequent commands:

grep -rl '\-\- MARK \-\-' /var/log/* | while read line; do echo "Working with file '${line}'"; grep MARK "$line" | tail -n1; done

Include lines before and after a grep match:

grep -B 3 -A 3 -i "hv_fcopy" /var/log/messages

Find all unique directories in listed directories that contain files modified 10 minutes ago since the command was ran:

ls | xargs -I {} find {} -type f -mmin -10 | cut -d "/" -f2 | sort -u

Find all files in the current directories that were modified at least a minute ago, are larger than 500MB, and long list them:

find . -type f -mmin -1 -size +500M -exec ls -lsh {} \;

Find all files in the current directories that were modified at least a day ago, are larger than 2GB, and empty their contents:

find . -type f -mtime -1 -size +2G -exec bash -c 'echo > {}' \;

Run arbitrary command against a list of directories:

ls | xargs -I {} git -C {} pull

Place the following somewhere in your shell script for easier step-by-step debugging; move ahead with Enter key:

set -x
trap read debug

Change timezone interactively:

dpkg-reconfigure tzdata

Search binary file that looks like text while ignoring case:

grep -ai "end:" /var/log/syslog

Count time, calls, and errors for each system call when performing a directory listing:

strace -c ls test/

Add to script to determine which line number the execution is at:

echo "DEBUG: ${LINENO}"

Remove duplicated lines from a file without messing up the order:

awk '!visited[$0]++' your_file > deduplicated_file

Run local script on a remote endpoint using SSH:

ssh -q <username>@<endpoint> "sudo bash -s" < local_script.sh

Create new directory and change right into it:

# Oneliner
mkdir new_directory && cd $_

# Alias
# https://unix.stackexchange.com/a/9124
mkcd () {
  mkdir "$1"
  cd "$1"
}

Recall argument to last used command:

$_
!$
Alt + .
# https://stackoverflow.com/a/3371711
!:1
!:1-2

Get SSH key fingerprint:

# SHA-256
ssh-keygen -lf ~/.ssh/id_rsa.pub

# MD5
ssh-keygen -E md5 -lf ~/.ssh/id_rsa.pub

Find broken symbolic links in current directory:

find . -xtype l

Bulk fix relative symbolic links:

find . -lname '<relative-to-source target>*' -exec sh -c 'ln -sfn "<new relative-to-source target>/$(basename $0)" $0' {} \;

Run remote script on remote endpoint using SSH:

ssh -q <username>@<endpoint> './location/to/script'

Run local script on remote endpoint using SSH:

ssh -q <username>@<endpoint> 'sudo bash -s' < ./location/to/local/script

Create ISO from directory without truncating long names (-l) and by not replacing hyphens with underscores -iso-level 4:

genisoimage -o data.iso -iso-level 4 -R -l data/

List ISO file contents without having to mount it:

isoinfo -l -i data.iso

Simple colouring for log files, both static and running output:

# https://automationrhapsody.com/coloured-log-files-linux/
cat test.log | perl -pe 's/^\[\*\].*/\e[0;36m$&\e[0m/g; s/^\[\+\].*/\e[0;32m$&\e[0m/g; s/^\[\!\].*/\e[0;31m$&\e[0m/g'

In situations such as these a Bash environment variable will suppress warnings that clog output:

export PYTHONWARNINGS='ignore'

Remove last column in string based on delimiter:

$ string='my_underscored_string_12345'
$ echo "$string" | rev | cut -d '_' -f 2- | rev
my_underscored_string

Prefix aliased command with backslash to avoid triggering alias:

$ halt -p
REALLY!? -p
$ alias halt
alias halt='echo "REALLY!?"'
$ \halt -p
Connection to example.com closed by remote host.

Pretty print CSV files:

# https://www.stefaanlippens.net/pretty-csv.html
function pretty_csv {
    perl -pe 's/((?<=,)|(?<=^)),/ ,/g;' "$@" | column -t -s, | less  -F -S -X -K
}

$ pretty_csv data.csv
$ pretty_csv < data.csv
$ sort data.csv | pretty_csv

Pretty print TSV files:

function pretty_tsv {
    perl -pe 's/((?<=\t)|(?<=^))\t/ \t/g;' "$@" | column -t -s $'\t' | less  -F -S -X -K
}

$ pretty_tsv data.tsv
$ pretty_tsv < data.tsv
$ sort data.tsv | pretty_tsv

Diff two files and save unified output to file:

diff -u file1 file2 > files.diff

Show build information for cloud-based image:

$ cat /etc/cloud/build.info 
build_name: server
serial: 20201211.1

Show top disk usage and exclude certain directories under root:

du -Sh / --exclude=/{proc,sys,dev,var} | sort -rh | head -n 10

It's possible to use the built-in : as a short-hand for an infinite loop:

while :; do "looping"; done

Re-execute a shell to 'unsource' variables and aliases:

exec /bin/bash

Use binary version of time instead of shell built-in to get access to more information:

# https://stackoverflow.com/questions/9006596/is-the-unix-time-command-accurate-enough-for-benchmarks
$(which time) --verbose echo "test"

Use perf stat to easily perform repeated executions of a command and measure it in various ways:

perf stat --null --repeat 5 --table echo "test"

Change nested key value in an array of JSON objects:

.parameters.vmObjects.value |= map(if .vmName == "router" then .moduleSnapshot = "fixed" else . end)


Use indirect references to use dynamic variable names:

for host in "${hosts[@]}"; do
  declare "parent_disk_${host}=$parent_disk"
done

for host in "${hosts[@]}"; do
  parent_disk="parent_disk_${host}"
  echo "${!parent_disk}"
done

Test terminal's colors:

msgcat --color=test

Bulk rename files in place:

find . -type f -name '<file name>' -execdir mv {} "description.txt" \;

Encode with Base64 on a single line:

echo "text" | base64 -w 0

Convert PEM to single-line:

awk 'NF {sub(/\r/, ""); printf "%s\\n",$0;}' combined.pem

Install requirements for Poetry using existing requirements.txt:

cat requirements.txt | xargs poetry add

Direct standard output to a file in a directory that might not yet exist:

echo "something" | install -D /dev/stdin directory/file.txt

Delete files older than 1 year:

find /the/dir/to/start/in -type f -mtime +365 -ls -exec rm -f -- {} \;

View permissions as a tree

 # -p - permissions
 # -u - username/userid
 # -f - full path
 # -i - don't print indentation lines
 # -d - print directories only
tree -pufid

Bulk uninstall pip packages according to a wildcard

pip freeze | grep "azure*" | xargs -n 1 pip uninstall -y

Show transaction history for a package:

dnf history list <package>

Show information about specific transaction in history:

More on history: https://www.putorius.net/dnf-history.html


dnf history info <transaction ID>

List SystemD timers:

systemctl list-timers

Show execution of service tied to a timer of an identical name:

journalctl -u name.timer
journalctl -u name.service

Copy remote directory to local system:

scp -rCp <user>@<remote>:<remote path> <local path>
# OR (faster)
rsync -azvhP <user>@<remote>:<remote path> <local path>

Overwrite existing directory with contents from another:

rsync -av --delete ~/new/ ~/old

Count number of installed kernels:

$ sudo dnf list --installed kernel-core* | tail -n +2 | wc -l
9

Increase number of installed kernels in /etc/dnf/dnf.conf:

...
installonly_limit=10
...


Pin specific kernel version:

$ sudo dnf install python3-dnf-plugins-extras-versionlock
$ # List kernel packages
$ rpm -qa kernel
kernel-6.0.18-300.fc37.x86_64
kernel-6.1.7-200.fc37.x86_64
kernel-6.1.8-200.fc37.x86_64
$ sudo dnf versionlock add kernel-6.0.18-300.fc37.x86_64
Last metadata expiration check: 3:51:11 ago on E 30 jaan  2023 15:47:21.
Adding versionlock on: kernel-0:6.0.18-300.fc37.*
$ # Remove pin
$ sudo dnf versionlock delete kernel-6.0.18-300.fc37.x86_64
...

Undo ad hoc changes made to a SystemD service (e.g. systemd-resolved):

$ systemctl revert systemd-resolved.service
Removed "/etc/systemd/system/systemd-resolved.service.d/override.conf".
Removed "/etc/systemd/system/systemd-resolved.service.d".
$ systemctl restart systemd-resolved.service

Back up a file using brace expansion:

Equivalent to cp /etc/ssh/sshd_config /etc/ssh/sshd_config.bak


cp /etc/ssh/sshd_config{,.bak}

Restore a backed up file:

Equivalent to cp /etc/ssh/sshd_config.bak /etc/ssh/sshd_config


cp /etc/ssh/sshd_config{.bak,}
Vim


Write file opened as 'readonly':

:w !sudo tee "%"


Visual editing can be used to delete any number of lines:

Shift+V from point of cursor and press D key to delete


Search and replace:

:%s/search/replace/g 


Disable search highlight in Vim:

:noh


Move line on cursor up or down:

# One line up
:m -2
# One line down
:m +1

Git


Push existing repository to new remote:

git remote add <name of new remote> <HTTPS or SSH URL>
git push <name of new remote> master

Pretty-print branch graph:

git log --all --decorate --oneline --graph

Move master back X amount of commits:

git checkout master
git reset --hard <old_commit_id>
git push -f origin master

Replay changes on master to some other branch:

# Beware of blindly accepting any incoming changes in favor of your own
# https://demisx.github.io/git/rebase/2015/07/02/git-rebase-keep-my-branch-changes.html
git checkout master
git pull
git checkout different_branch
git rebase -Xtheirs master
git push --force

Show changed files in specified commit hash:

# https://stackoverflow.com/questions/49853177/how-to-see-which-files-were-changed-in-last-commit
git diff-tree --no-commit-id --name-only <commit hash>

Create patch file from diff:

git diff file.json > file.patch

Create patch file from commit:

git show <commit hash> > commit.patch

Apply patch file:

git apply commit.patch

Bulk create patch files from individual files when running git diff in a repository:

https://raymii.org/s/tutorials/Bash_bits_split_a_file_in_blocks_and_do_something_with_each_block.html
OLDIFS=$IFS; IFS=';' blocks=$(git diff | sed -n '/diff/,/(diff|$)/ {/diff / s/^/\;/; p}'); for block in ${blocks#;}; do echo "$block" > $(echo "$block" | head -n 1 | rev | cut -d "/" -f 1 | rev).patch; done; IFS=$OLDIFS

Show diff of stashed hunk:

git stash show -p [stash@{N}]

Bulk create separate stashes of every changed file with a message equaling the filename:

git status -s | cut -d " " -f 3 | xargs -I {} git stash push {} -m "{}"

Pop every entry from the stash back to the working tree:

git stash list | cut -d ":" -f 1 | xargs -I {} git stash pop

Move unpushed commits to a new branch:

# Pull latest changes from 'origin/master' if haven't already
# https://stackoverflow.com/a/46726955
git checkout -b new_branch
git checkout master
git reset --hard origin/master

Copy commit to current branch:

git cherry-pick <commit hash>

Undo pushed commit that nobody has yet pulled:

git reset HEAD^ --hard
git push --force origin

View history of specific function in file:

git log -L :<function>:<file>

Speed up Git for larger repositories:

git config feature.manyFiles 1

Search through history for a specific word:

git rev-list --all | ( while read revision; do git grep -F 'word' "$revision"; done; )

Delete remote branch:

git push origin --delete branch/name

Bulk reset author of multiple (unpushed) commits (e.g. 9):

# Set correct user name and email prior to this
git rebase --onto HEAD~9 --exec "git commit --amend --reset-author --no-edit" HEAD~9

Re-order commits:

git rebase --interactive
# Oldest commit will be at the top
# Move commit down with 'ddp'
# Move commit up with 'ddkP'

Search for 'something' in a commit message:

git log --all -i --grep='something'

Search for 'something' through all commits' contents:

git grep 'something' $(git rev-list --all)
Docker/Podman


Parse lines from an arguments file to separate parameters for building:

# https://ilhicas.com/2018/11/03/docker-build-with-build-arg-wit-multiple-arguments.html
podman build -t foo $(while IFS= read -r line; do args+="--build-arg ${line} "; done < <(cat .arg); echo "$args"; unset args) .

Remove all 'exited' containers:

podman rm $(podman ps --all -q -f status=exited)

Build and run container based on Dockerfile in current context:

podman build -t foo . && podman run --rm -it foo

Prune everything that shouldn't exist anymore without any confirmation:

podman system prune -a -f

Remove all images except latest:

podman images | grep -v "latest" | tail -n +2 | awk '{ print $3 }' | xargs --no-run-if-empty podman rmi

Possible improvement when executing RUN within a Dockerfile:

Benefit is that when a specific line fails, then the error message is much more concise as opposed to the standard method of using ampersands


RUN set -eu; \
    python3 -m venv venv; \
    venv/bin/pip install -r requirements.txt; \
    venv/bin/pip install -r requirements-dev.txt; \
    echo 'Venv creation + requirements installation: OK';


Remove dangling <none>:<none> images:

docker rmi $(docker images -f "dangling=true" -q)
Ansible


Run role against arbitrary host:

# https://stackoverflow.com/a/38384205
# Note the comma after the IP or FQDN
# https://groups.google.com/d/msg/ansible-project/G_9JRGp5jGE/PTBZdgDb5OEJ
# Additional hosts can be added by supplying '-i' parameter with more arguments (comma at the end only if count == 1)
ansible-playbook -i '<IP or FQDN>,' -u '<user name>' --extra-vars 'ansible_winrm_server_cert_validation=ignore ansible_connection=winrm ansible_winrm_transport=credssp ansible_password=<password>' --tags '<tag value for a role>' playbook.yml

Run ad-hoc command against arbitrary host:

# Replace the final 'all' with a more precise host pattern if you passed more than one IP or FQDN to the initial list (comma at the end only if count == 1)
ansible -i '<IP or FQDN>,' -u '<user name>' --extra-vars 'ansible_winrm_server_cert_validation=ignore ansible_connection=winrm ansible_winrm_transport=credssp ansible_password=<password>' -m 'win_shell' -a 'ipconfig' 'all'

Add timing information to playbook execution output:

# https://docs.ansible.com/ansible/latest/plugins/callback/profile_tasks.html
$ cat ansible.cfg
[defaults]
callback_whitelist = profile_tasks

Make verbose output more readable by using YAML instead of JSON:

ANSIBLE_STDOUT_CALLBACK='yaml'

Debug variables without running entire playbook:

ansible -m debug <host> -a "var=hostvars[inventory_hostname].<variable>"
Grafana/Promtail


Test simple journal scraping configuration:

$ cat /etc/promtail/config.yml
server:
  http_listen_port: 9080
  grpc_listen_port: 0

positions:
  filename: /var/log/positions.yaml

clients:
  - url: https://<token>@<host>/loki/api/v1/push

scrape_configs:
  - job_name: journal
    journal:
      max_age: 12h
      labels:
        job: systemd-journal
        static_label: label-value
    relabel_configs:
      - source_labels: ['__journal__systemd_unit']
        target_label: 'unit'

$ promtail --dry-run --config.file /etc/promtail/config.yml
2021-05-20T03:48:54     {job="systemd-journal", static_label="label-value"}     Stopped target Default.
2021-05-20T03:48:54     {job="systemd-journal", static_label="label-value"}     Stopped target Basic System.
2021-05-20T03:48:54     {job="systemd-journal", static_label="label-value"}     Stopped target Timers.
2021-05-20T03:48:54     {job="systemd-journal", static_label="label-value"}     Stopped target Paths.
2021-05-20T03:48:54     {job="systemd-journal", static_label="label-value"}     Stopped target Sockets.
PowerShell


Get today's date and time where time is set to midnight:

[DateTime]::Today

Show list of logged on users:

query user /server:$SERVER

Log off user by specifying session ID:

logoff <session ID>

Reload a local module:

Import-Module -Name .\module.psd1 -Force

Pretty-print minified JSON:

$String | ConvertFrom-Json | ConvertTo-Json -Depth 100

Convert from Base64:

[Text.Encoding]::Utf8.GetString([Convert]::FromBase64String($String))

Convert string to boolean:

[System.Convert]::ToBoolean($String)

Resolve FQDN:

[System.Net.Dns]::GetHostByName($FQDN)
Windows


Colored prompt with Git branch for terminal (e.g. ConEmu):

PS1='${debian_chroot:+($debian_chroot)}\[\033[01;35m\] \w\[\033[01;33m\]$(__git_ps1)\[\033[01;m\] > '

Miscellaneous


Query in JMESPath website: [?tags.currently-used == 'False' || tags.currently_used == 'False'].name

Same query in jpterm: [?tags."currently-used" == 'False' || tags."currently_used" == 'False']
Same query in console: "[?tags.\"currently-used\" == 'False' || tags.\"currently_used\" == 'False'].name"


Convert x265 MP4 to x264: ffmpeg -i input.mp4 -c:v libx264 -crf 20 -c:a copy output.mp4
Create Windows install media:

Download WoeUSB
Download Windows 10 ISO
Find out USB's disk: sudo fdisk --list
Wipe disk entirely and burn ISO: sudo ./woeusb-5.2.4.bash --target-filesystem NTFS --device Win10_22H2_EnglishInternational_x64.iso /dev/sda