Linux Snippets

Send logs of a cronjob to journal⚑

You can use systemd-cat to send the logs of a script or cron to the journal to the unit specified after the -t flag. It works better than piping the output to logger -t

systemd-cat -t syncoid_send_backups /root/send_backups.sh

Set dependencies between systemd services⚑

Use Wants or Requires:

website.service
[Unit]
Wants=mongodb.service
After=mongodb.service

Set environment variable in systemd service⚑

[Service]
# ...
Environment="FOO=foo"

Get info of a mkv file⚑

ffprobe file.mkv

Send multiline messages with notify-send⚑

The title can't have new lines, but the body can.

notify-send "Title" "This is the first line.\nAnd this is the second.")

Find BIOS version⚑

dmidecode | less

Reboot server on kernel panic ⚑

The proc/sys/kernel/panic file gives read/write access to the kernel variable panic_timeout. If this is zero, the kernel will loop on a panic; if nonzero it indicates that the kernel should autoreboot after this number of seconds. When you use the software watchdog device driver, the recommended setting is 60.

To set the value add the next contents to the /etc/sysctl.d/99-panic.conf

kernel.panic = 60

Or with an ansible task:

- name: Configure reboot on kernel panic
  become: true
  lineinfile:
    path: /etc/sysctl.d/99-panic.conf
    line: kernel.panic = 60
    create: true
    state: present

If you want the machine to automatically reboot, make sure you set kernel.panic to something above 0. Otherwise these settings could cause a hung machine that you will have to reboot manually.

# Panic if the kernel detects an I/O channel
# check (IOCHK). 0=no | 1=yes
kernel.panic_on_io_nmi = 1

# Panic if a hung task was found. 0=no, 1=yes
kernel.hung_task_panic = 1

# Setup timeout for hung task,
# in seconds (suggested 300)
kernel.hung_task_timeout_secs = 300

# Panic on out of memory.
# 0=no | 1=usually | 2=always
vm.panic_on_oom=2

# Panic when the kernel detects an NMI
# that usually indicates an uncorrectable
# parity or ECC memory error. 0=no | 1=yes
kernel.panic_on_unrecovered_nmi=1

# Panic if the kernel detects a soft-lockup
# error (1). Otherwise it lets the watchdog
# process skip it's update (0)
# kernel.softlockup_panic=0

# Panic on oops too. Use with caution.
# kernel.panic_on_oops=30

Share a calculated value between github actions steps⚑

You need to set a step's output parameter. Note that the step will need an id to be defined to later retrieve the output value.

echo "{name}={value}" >> "$GITHUB_OUTPUT"

For example:

- name: Set color
  id: color-selector
  run: echo "SELECTED_COLOR=green" >> "$GITHUB_OUTPUT"
- name: Get color
  env:
    SELECTED_COLOR: ${{ steps.color-selector.outputs.SELECTED_COLOR }}
  run: echo "The selected color is $SELECTED_COLOR"

Split a zip into sizes with restricted size ⚑

Something like:

zip -9 myfile.zip *
zipsplit -n 250000000 myfile.zip

Would produce myfile1.zip, myfile2.zip, etc., all independent of each other, and none larger than 250MB (in powers of ten). zipsplit will even try to organize the contents so that each resulting archive is as close as possible to the maximum size.

find files that were modified between dates⚑

The best option is the -newerXY. The m and t flags can be used.

• m The modification time of the file reference
• t reference is interpreted directly as a time

So the solution is

find . -type f -newermt 20111222 \! -newermt 20111225

The lower bound in inclusive, and upper bound is exclusive, so I added 1 day to it. And it is recursive.

Rotate image with the command line ⚑

If you want to overwrite in-place, mogrify from the ImageMagick suite seems to be the easiest way to achieve this:

# counterclockwise:
mogrify -rotate -90 *.jpg

# clockwise:
mogrify -rotate 90 *.jpg

Configure desktop icons in gnome⚑

Latest versions of gnome dont have desktop icons read this article to fix this

Make a file executable in a git repository ⚑

git add entrypoint.sh
git update-index --chmod=+x entrypoint.sh

Configure autologin in Debian with Gnome⚑

Edit the /etc/gdm3/daemon.conf file and include:

AutomaticLoginEnable = true
AutomaticLogin = <your user>

See errors in the journalctl ⚑

To get all errors for running services using journalctl:

journalctl -p 3 -xb

Fix rsyslog builtin:omfile suspended error⚑

It may be a permissions error. I have not been able to pinpoint the reason behind it.

What did solve it though is to remove the aledgely deprecated paramenters from /etc/rsyslog.conf:

# $FileOwner syslog
# $FileGroup adm
# $FileCreateMode 0640
# $DirCreateMode 0755 
# $Umask 0022 
# $PrivDropToUser syslog 
# $PrivDropToGroup syslog 

I hope that as they are the default parameters, they don't need to be set.

Configure nginx to restrict methods⚑

server {
    listen 80;
    server_name yourdomain.com;

    location / {
        if ($request_method !~ ^(GET|POST)$ ) {
            return 405;
        }

        try_files $uri $uri/ =404;
    }
}

Configure nginx location regexp to accept dashes⚑

location ~* /share/[\w-]+ {
  root /home/project_root;
}

Configure nginx location to accept many paths⚑

location ~ ^/(static|media)/ {
  root /home/project_root;
}

Remove image metadata⚑

exiftool -all:all= /path/to/file

Get the size increment of a directory between two dates⚑

To see how much has a directory grown between two dates you can use:

find /path/to/directory -type f -newerat 2022-12-31 ! -newerat 2024-01-01 -printf "%s\\n" | awk '{s+=$1} END {print s}'

It finds all the files in that directory that were created in the 2023, it only prints their size in bytes and then it adds them all up.

Makefile use bash instead of sh⚑

The program used as the shell is taken from the variable SHELL. If this variable is not set in your makefile, the program /bin/sh is used as the shell.

So put SHELL := /bin/bash at the top of your makefile, and you should be good to go.

Recover the message of a commit if the command failed⚑

git commit can fail for reasons such as gpg.commitsign = true && gpg fails, or when running a pre-commit. Retrying the command opens a blank editor and the message seems to be lost.

The message is saved though in .git/COMMIT_EDITMSG, so you can:

git commit -m "$(cat .git/COMMIT_EDITMSG)"

Or in general (suitable for an alias for example):

git commit -m "$(cat "$(git rev-parse --git-dir)/COMMIT_EDITMSG)")"

Accept new ssh keys by default⚑

While common wisdom is not to disable host key checking, there is a built-in option in SSH itself to do this. It is relatively unknown, since it's new (added in Openssh 6.5).

This is done with -o StrictHostKeyChecking=accept-new. Or if you want to use it for all hosts you can add the next lines to your ~/.ssh/config:

Host *
  StrictHostKeyChecking accept-new

WARNING: use this only if you absolutely trust the IP\hostname you are going to SSH to:

ssh -o StrictHostKeyChecking=accept-new mynewserver.example.com

Note, StrictHostKeyChecking=no will add the public key to ~/.ssh/known_hosts even if the key was changed. accept-new is only for new hosts. From the man page:

If this flag is set to “accept-new” then ssh will automatically add new host keys to the user known hosts files, but will not permit connections to hosts with changed host keys. If this flag is set to “no” or “off”, ssh will automatically add new host keys to the user known hosts files and allow connections to hosts with changed hostkeys to proceed, subject to some restrictions. If this flag is set to ask (the default), new host keys will be added to the user known host files only after the user has confirmed that is what they really want to do, and ssh will refuse to connect to hosts whose host key has changed. The host keys of known hosts will be verified automatically in all cases.

Do not add trailing / to ls⚑

Probably, your ls is aliased or defined as a function in your config files.

Use the full path to ls like:

/bin/ls /var/lib/mysql/ 

Convert png to svg⚑

Inkscape has got an awesome auto-tracing tool.

• Install Inkscape using sudo apt-get install inkscape
• Import your image
• Select your image
• From the menu bar, select Path > Trace Bitmap Item
• Adjust the tracing parameters as needed
• Save as svg

Check their tracing tutorial for more information.

Once you are comfortable with the tracing options. You can automate it by using CLI of Inkscape.

Redirect stdout and stderr of a cron job to a file⚑

*/1 * * * * /home/ranveer/vimbackup.sh >> /home/ranveer/vimbackup.log 2>&1

Error when unmounting a device: Target is busy⚑

• Check the processes that are using the mountpoint with lsof /path/to/mountpoint
• Kill those processes
• Try the umount again

If that fails, you can use umount -l.

Wipe a disk⚑

Overwrite it many times with badblocks.

badblocks -wsv -b 4096 /dev/sde | tee disk_wipe_log.txt

Impose load on a system to stress it⚑

sudo apt-get install stress

stress --cpu 2

That will fill up the usage of 2 cpus. To run 1 vm stressor using 1GB of virtual memory for 60s, enter:

stress --vm 1 --vm-bytes 1G --vm-keep -t 60s

You can also stress io with --io 4, for example to spawn 4 workers.

Get the latest tag of a git repository⚑

git describe --tags --abbrev=0

Configure gpg-agent cache ttl⚑

The user configuration (in ~/.gnupg/gpg-agent.conf) can only define the default and maximum caching duration; it can't be disabled.

The default-cache-ttl option sets the timeout (in seconds) after the last GnuPG activity (so it resets if you use it), the max-cache-ttl option set the timespan (in seconds) it caches after entering your password. The default value is 600 seconds (10 minutes) for default-cache-ttl and 7200 seconds (2 hours) for max-cache-ttl.

default-cache-ttl 21600
max-cache-ttl 21600

For this change to take effect, you need to end the session by restarting gpg-agent.

gpgconf --kill gpg-agent
gpg-agent --daemon --use-standard-socket

Get return code of failing find exec⚑

When you run find . -exec ls {} \; even if the command run in the exec returns a status code different than 0 you'll get an overall status code of 0 which makes difficult to catch errors in bash scripts.

You can instead use xargs, for example:

find /tmp/ -iname '*.sh' -print0 | xargs -0 shellcheck

This will run shellcheck file_name for each of the files found by the find command.

Limit the resources a docker is using⚑

You can either use limits in the docker service itself, see 1 and 2.

Or/and you can limit it for each docker, see 1 and 2.

Get the current git branch⚑

git branch --show-current

Install latest version of package from backports⚑

Add the backports repository:

vi /etc/apt/sources.list.d/bullseye-backports.list

deb http://deb.debian.org/debian bullseye-backports main contrib
deb-src http://deb.debian.org/debian bullseye-backports main contrib

Configure the package to be pulled from backports

vi /etc/apt/preferences.d/90_zfs

Package: src:zfs-linux
Pin: release n=bullseye-backports
Pin-Priority: 990

Rename multiple files matching a pattern⚑

There is rename that looks nice, but you need to install it. Using only find you can do:

find . -name '*yml' -exec bash -c 'echo mv $0 ${0/yml/yaml}' {} \; 

If it shows what you expect, remove the echo.

Force ssh to use password authentication⚑

ssh -o PreferredAuthentications=password -o PubkeyAuthentication=no exampleUser@example.com

Do a tail -f with grep⚑

tail -f file | grep --line-buffered my_pattern

Check if a program exists in the user's PATH⚑

command -v <the_command>

Example use:

if ! command -v <the_command> &> /dev/null
then
    echo "<the_command> could not be found"
    exit
fi

Reset failed systemd services⚑

Use systemctl to remove the failed status. To reset all units with failed status:

systemctl reset-failed

or just your specific unit:

systemctl reset-failed openvpn-server@intranert.service

Automatic reboot after power failure⚑

That's not something you can control in your operating system. That's what the BIOS is for. In most BIOS setups there'll be an option like After power loss with possible values like Power off and Reboot.

You can also edit /etc/default/grub and add:

GRUB_RECORDFAIL_TIMEOUT=5

Then run:

sudo update-grub

This will make your machine display the boot options for 5 seconds before it boot the default option (instead of waiting forever for you to choose one).

SSH tunnel⚑

ssh -D 9090 -N -f user@host

If you need to forward an external port to a local one you can use

ssh -L LOCAL_PORT:DESTINATION:DESTINATION_PORT [USER@]SSH_SERVER

If you need a more powerful solution you can try sshuttle

Fix the SSH client kex_exchange_identification: read: Connection reset by peer error⚑

Restart the ssh service.

Get class of a window⚑

Use xprop and click the window.

Change the brightness of the screen⚑

Get the current brightness level with cat /sys/class/backlight/intel_backlight/brightness. Imagine it's 1550, then if you want to lower the brightness use:

sudo echo 500 > /sys/class/backlight/intel_backlight/brightness

Force umount nfs mounted directory⚑

umount -l path/to/mounted/dir

Configure fstab to mount nfs⚑

NFS stands for ‘Network File System’. This mechanism allows Unix machines to share files and directories over the network. Using this feature, a Linux machine can mount a remote directory (residing in a NFS server machine) just like a local directory and can access files from it.

An NFS share can be mounted on a machine by adding a line to the /etc/fstab file.

The default syntax for fstab entry of NFS mounts is as follows.

Server:/path/to/export /local_mountpoint nfs <options> 0 0

Where:

• Server: The hostname or IP address of the NFS server where the exported directory resides.
• /path/to/export: The shared directory (exported folder) path.
• /local_mountpoint: Existing directory in the host where you want to mount the NFS share.

You can specify a number of options that you want to set on the NFS mount:

• soft/hard: When the mount option hard is set, if the NFS server crashes or becomes unresponsive, the NFS requests will be retried indefinitely. You can set the mount option intr, so that the process can be interrupted. When the NFS server comes back online, the process can be continued from where it was while the server became unresponsive.

When the option soft is set, the process will be reported an error when the NFS server is unresponsive after waiting for a period of time (defined by the timeo option). In certain cases soft option can cause data corruption and loss of data. So, it is recommended to use hard and intr options.

• noexec: Prevents execution of binaries on mounted file systems. This is useful if the system is mounting a non-Linux file system via NFS containing incompatible binaries.
• nosuid: Disables set-user-identifier or set-group-identifier bits. This prevents remote users from gaining higher privileges by running a setuid program.
• tcp: Specifies the NFS mount to use the TCP protocol.
• udp: Specifies the NFS mount to use the UDP protocol.
• nofail: Prevent issues when rebooting the host. The downside is that if you have services that depend on the volume to be mounted they won't behave as expected.

Fix limit on the number of inotify watches⚑

Programs that sync files such as dropbox, git etc use inotify to notice changes to the file system. The limit can be see by -

cat /proc/sys/fs/inotify/max_user_watches

For me, it shows 65536. When this limit is not enough to monitor all files inside a directory it throws this error.

If you want to increase the amount of inotify watchers, run the following in a terminal:

echo fs.inotify.max_user_watches=100000 | sudo tee -a /etc/sysctl.conf && sudo sysctl -p

Where 100000 is the desired number of inotify watches.

What is /var/log/tallylog⚑

/var/log/tallylog is the file where the PAM linux module (used for authentication of the machine) keeps track of the failed ssh logins in order to temporarily block users.

Manage users⚑

• Change main group of user

usermod -g {{ group_name }} {{ user_name }}

• Add user to group

usermod -a -G {{ group_name }} {{ user_name }}

• Remove user from group.

usermod -G {{ remaining_group_names }} {{ user_name }}

You have to execute groups {{ user }} get the list and pass the remaining to the above command

• Change uid and gid of the user

usermod -u {{ newuid }} {{ login }}
groupmod -g {{ newgid }} {{ group }}
find / -user {{ olduid }} -exec chown -h {{ newuid }} {} \;
find / -group {{ oldgid }} -exec chgrp -h {{ newgid }} {} \;
usermod -g {{ newgid }} {{ login }}

Manage ssh keys⚑

• Generate ed25519 key

ssh-keygen -t ed25519 -f {{ path_to_keyfile }}

• Generate RSA key

ssh-keygen -t rsa -b 4096 -o -a 100 -f {{ path_to_keyfile }}

• Generate different comment

ssh-keygen -t ed25519 -f {{ path_to_keyfile }} -C {{ email }}

• Generate key headless, batch

ssh-keygen -t ed25519 -f {{ path_to_keyfile }} -q -N ""

• Generate public key from private key

ssh-keygen -y -f {{ path_to_keyfile }} > {{ path_to_public_key_file }}

• Get fingerprint of key

  ssh-keygen -lf {{ path_to_key }}
  

Measure the network performance between two machines⚑

Install iperf3 with apt-get install iperf3 on both server and client.

On the server system run:

server#: iperf3 -i 10 -s

Where:

• -i: the interval to provide periodic bandwidth updates
• -s: listen as a server

On the client system:

client#: iperf3 -i 10 -w 1M -t 60 -c [server hostname or ip address]

Where:

• -i: the interval to provide periodic bandwidth updates
• -w: the socket buffer size (which affects the TCP Window). The buffer size is also set on the server by this client command.
• -t: the time to run the test in seconds
• -c: connect to a listening server at…

Sometimes is interesting to test both ways as they may return different outcomes

I've got the next results at home:

• From new NAS to laptop through wifi 67.5 MB/s
• From laptop to new NAS 59.25 MB/s
• From intel Nuc to new NAS 116.75 MB/s (934Mbit/s)
• From old NAS to new NAS 11 MB/s

Measure the performance, IOPS of a disk⚑

To measure disk IOPS performance in Linux, you can use the fio tool. Install it with

apt-get install fio

Then you need to go to the directory where your disk is mounted. The test is done by performing read/write operations in this directory.

To do a random read/write operation test an 8 GB file will be created. Then fio will read/write a 4KB block (a standard block size) with the 75/25% by the number of reads and writes operations and measure the performance.

fio --randrepeat=1 --ioengine=libaio --direct=1 --gtod_reduce=1 --name=fiotest --filename=testfio --bs=4k --iodepth=64 --size=8G --readwrite=randrw --rwmixread=75

I've run this test in different environments with awesome results:

• New NAS server NVME:

  read: IOPS=297k, BW=1159MiB/s (1215MB/s)(3070MiB/2649msec)
   bw (  MiB/s): min= 1096, max= 1197, per=99.80%, avg=1156.61, stdev=45.31, samples=5
   iops        : min=280708, max=306542, avg=296092.00, stdev=11598.11, samples=5
  write: IOPS=99.2k, BW=387MiB/s (406MB/s)(1026MiB/2649msec); 0 zone resets
   bw (  KiB/s): min=373600, max=408136, per=99.91%, avg=396248.00, stdev=15836.85, samples=5
   iops        : min=93400, max=102034, avg=99062.00, stdev=3959.21, samples=5
  cpu          : usr=15.67%, sys=67.18%, ctx=233314, majf=0, minf=8
  

• New NAS server ZFS pool with RAIDZ:

read: IOPS=271k, BW=1059MiB/s (1111MB/s)(3070MiB/2898msec)
 bw (  MiB/s): min=  490, max= 1205, per=98.05%, avg=1038.65, stdev=306.74, samples=5
 iops        : min=125672, max=308484, avg=265893.20, stdev=78526.52, samples=5
write: IOPS=90.6k, BW=354MiB/s (371MB/s)(1026MiB/2898msec); 0 zone resets
 bw (  KiB/s): min=167168, max=411776, per=98.26%, avg=356236.80, stdev=105826.16, samples=5
 iops        : min=41792, max=102944, avg=89059.20, stdev=26456.54, samples=5
cpu          : usr=12.84%, sys=63.20%, ctx=234345, majf=0, minf=6

• Laptop NVME:

read: IOPS=36.8k, BW=144MiB/s (151MB/s)(3070MiB/21357msec)
 bw (  KiB/s): min=129368, max=160304, per=100.00%, avg=147315.43, stdev=6640.40, samples=42
 iops        : min=32342, max=40076, avg=36828.86, stdev=1660.10, samples=42
write: IOPS=12.3k, BW=48.0MiB/s (50.4MB/s)(1026MiB/21357msec); 0 zone resets
 bw (  KiB/s): min=42952, max=53376, per=100.00%, avg=49241.33, stdev=2151.40, samples=42
 iops        : min=10738, max=13344, avg=12310.33, stdev=537.85, samples=42
cpu          : usr=14.32%, sys=32.17%, ctx=356674, majf=0, minf=7

• Laptop ZFS pool through NFS (running in parallel with other network processes):

read: IOPS=4917, BW=19.2MiB/s (20.1MB/s)(3070MiB/159812msec)
 bw (  KiB/s): min=16304, max=22368, per=100.00%, avg=19681.46, stdev=951.52, samples=319
 iops        : min= 4076, max= 5592, avg=4920.34, stdev=237.87, samples=319
write: IOPS=1643, BW=6574KiB/s (6732kB/s)(1026MiB/159812msec); 0 zone resets
 bw (  KiB/s): min= 5288, max= 7560, per=100.00%, avg=6577.35, stdev=363.32, samples=319
 iops        : min= 1322, max= 1890, avg=1644.32, stdev=90.82, samples=319
cpu          : usr=5.21%, sys=10.59%, ctx=175825, majf=0, minf=8

• Intel Nuc server disk SSD:

    read: IOPS=11.0k, BW=46.9MiB/s (49.1MB/s)(3070MiB/65525msec)
   bw (  KiB/s): min=  280, max=73504, per=100.00%, avg=48332.30, stdev=25165.49, samples=130
   iops        : min=   70, max=18376, avg=12083.04, stdev=6291.41, samples=130
  write: IOPS=4008, BW=15.7MiB/s (16.4MB/s)(1026MiB/65525msec); 0 zone resets
   bw (  KiB/s): min=   55, max=24176, per=100.00%, avg=16153.84, stdev=8405.53, samples=130
   iops        : min=   13, max= 6044, avg=4038.40, stdev=2101.42, samples=130
  cpu          : usr=8.04%, sys=25.87%, ctx=268055, majf=0, minf=8
  

• Intel Nuc server external HD usb disk :

• Intel Nuc ZFS pool through NFS (running in parallel with other network processes):

  read: IOPS=18.7k, BW=73.2MiB/s (76.8MB/s)(3070MiB/41929msec)
 bw (  KiB/s): min=43008, max=103504, per=99.80%, avg=74822.75, stdev=16708.40, samples=83
 iops        : min=10752, max=25876, avg=18705.65, stdev=4177.11, samples=83
write: IOPS=6264, BW=24.5MiB/s (25.7MB/s)(1026MiB/41929msec); 0 zone resets
 bw (  KiB/s): min=14312, max=35216, per=99.79%, avg=25003.55, stdev=5585.54, samples=83
 iops        : min= 3578, max= 8804, avg=6250.88, stdev=1396.40, samples=83
cpu          : usr=6.29%, sys=13.21%, ctx=575927, majf=0, minf=10

• Old NAS with RAID5:

  read : io=785812KB, bw=405434B/s, iops=98, runt=1984714msec
  write: io=262764KB, bw=135571B/s, iops=33, runt=1984714msec
  cpu          : usr=0.16%, sys=0.59%, ctx=212447, majf=0, minf=8
  

Conclusions:

• New NVME are super fast (1215MB/s read, 406MB/s write)
• ZFS rocks, with a RAIDZ1, L2ARC and ZLOG it returned almost the same performance as the NVME ( 1111MB/s read, 371MB/s write)
• Old NAS with RAID is super slow (0.4KB/s read, 0.1KB/s write!)
• I should replace the laptop's NVME, the NAS one has 10x performace both on read and write.

There is a huge difference between ZFS in local and through NFS. In local you get (1111MB/s read and 371MB/s write) while through NFS I got (20.1MB/s read and 6.7MB/s write). I've measured the network performance between both machines with iperf3 and got:

• From NAS to laptop 67.5 MB/s
• From laptop to NAS 59.25 MB/s

It was because I was running it over wifi.

From the Intel nuc to the new server I get 76MB/s read and 25.7MB/s write. Still a huge difference though against the local transfer. The network speed measured with iperf3 are 116 MB/s.

Use a pass password in a Makefile⚑

TOKEN ?= $(shell bash -c '/usr/bin/pass show path/to/token')

diff:
    @AUTHENTIK_TOKEN=$(TOKEN) terraform plan

Install a new font⚑

Install a font manually by downloading the appropriate .ttf or otf files and placing them into /usr/local/share/fonts (system-wide), ~/.local/share/fonts (user-specific) or ~/.fonts (user-specific). These files should have the permission 644 (-rw-r--r--), otherwise they may not be usable.

Get VPN password from pass⚑

To be able to retrieve the user and password from pass you need to run the openvpn command with the next flags:

sudo bash -c "openvpn --config config.ovpn  --auth-user-pass <(echo -e 'user_name\n$(pass show vpn)')"

Assuming that vpn is an entry of your pass password store.

Download TS streams⚑

Some sites give stream content with small .ts files that you can't download directly. Instead open the developer tools, reload the page and search for a request with extension .m3u8, that gives you the playlist of all the chunks of .ts files. Once you have that url you can use yt-dlp to download it.

df and du showing different results⚑

Sometimes on a linux machine you will notice that both df command (display free disk space) and du command (display disk usage statistics) report different output. Usually, df will output a bigger disk usage than du.

The du command estimates file space usage, and the df command shows file system disk space usage.

There are many reasons why this could be happening:

Disk mounted over data⚑

If you mount a disk on a directory that already holds data, then when you run du that data won't show, but df knows it's there.

To troubleshoot this, umount one by one of your disks, and do an ls to see if there's any remaining data in the mount point.

Used deleted files⚑

When a file is deleted under Unix/Linux, the disk space occupied by the file will not be released immediately in some cases. The result of the command du doesn’t include the size of the deleting file. But the impact of the command df for the deleting file’s size due to its disk space is not released immediately. Hence, after deleting the file, the results of df and du are different until the disk space is freed.

Open file descriptor is main causes of such wrong information. For example, if a file called /tmp/application.log is open by a third-party application OR by a user and the same file is deleted, both df and du report different outputs. You can use the lsof command to verify this:

lsof | grep tmp

To fix it:

• Use the lsof command as discussed above to find a deleted file opened by other users and apps. See how to list all users in the system for more info.
• Then, close those apps and log out of those Linux and Unix users.
• As a sysadmin you restart any process or kill the process under Linux and Unix that did not release the deleted file.
• Flush the filesystem using the sync command that synchronizes cached writes to persistent disk storage.
• If everything else fails, try restarting the system using the reboot command or shutdown command.

Scan a physical page in Linux⚑

Install xsane and run it.

Git checkout to main with master as a fallback⚑

I usually use the alias gcm to change to the main branch of the repository, given the change from main to master now I have some repos that use one or the other, but I still want gcm to go to the correct one. The solution is to use:

alias gcm='git checkout "$(git symbolic-ref refs/remotes/origin/HEAD | cut -d'/' -f4)"'

Create QR code⚑

qrencode -o qrcode.png 'Hello World!'

Trim silences of sound files⚑

To trim all silence longer than 2 seconds down to only 2 seconds long.

sox in.wav out6.wav silence -l 1 0.1 1% -1 2.0 1%

Note that SoX does nothing to bits of silence shorter than 2 seconds.

If you encounter the sox FAIL formats: no handler for file extension 'mp3' error you'll need to install the libsox-fmt-all package.

Adjust the replay gain of many sound files⚑

sudo apt-get install python-rgain
replaygain -f *.mp3

Check vulnerabilities in Node.js applications⚑

With yarn audit you'll see the vulnerabilities, with yarn outdated you can see the packages that you need to update.

Check vulnerabilities in rails dependencies⚑

gem install bundler-audit
cd project_with_gem_lock
bundler-audit

Create Basic Auth header⚑

$ echo -n user:password | base64
dXNlcjpwYXNzd29yZA==

Without the -n it won't work well.

Install one package from Debian unstable⚑

• Add the unstable repository to your /etc/apt/sources.list

# Unstable
deb http://deb.debian.org/debian/ unstable main contrib non-free
deb-src http://deb.debian.org/debian/ unstable main contrib non-free

• Configure apt to only use unstable when specified

  File: /etc/apt/preferences

  Package: * 
  Pin: release a=stable
  Pin-Priority: 700
  
  Package: *
  Pin: release  a=testing
  Pin-Priority: 600
  
  Package: *
  Pin: release a=unstable
  Pin-Priority: 100
  

• Update the package data with apt-get update.

• See that the new versions are available with apt-cache policy <package_name>
• To install a package from unstable you can run apt-get install -t unstable <package_name>.

Fix the following packages have been kept back⚑

sudo apt-get --with-new-pkgs upgrade

Monitor outgoing traffic⚑

Easy and quick way watch & lsof⚑

You can simply use a combination of watch & lsof command in Linux to get an idea of outgoing traffic on specific ports. Here is an example of outgoing traffic on ports 80 and 443.

$ watch -n1 lsof -i TCP:80,443

Here is a sample output.

dropbox    2280 saml   23u  IPv4 56015285      0t0  TCP www.example.local:56003->snt-re3-6c.sjc.dropbox.com:http (ESTABLISHED)
thunderbi  2306 saml   60u  IPv4 56093767      0t0  TCP www.example.local:34788->ord08s09-in-f20.1e100.net:https (ESTABLISHED)
mono       2322 saml   15u  IPv4 56012349      0t0  TCP www.example.local:54018->204-62-14-135.static.6sync.net:https (ESTABLISHED)
chrome    4068 saml  175u  IPv4 56021419      0t0  TCP www.example.local:42182->stackoverflow.com:http (ESTABLISHED)

You'll miss the short lived connections though.

Fine grained with tcpdump⚑

You can also use tcpdump command to capture all raw packets, on all interfaces, on all ports, and write them to file.

sudo tcpdump -tttt -i any -w /tmp/http.log

Or you can limit it to a specific port adding the arguments port 443 or 80. The -tttt flag is used to capture the packets with a human readable timestamp.

To read the recorded information, run the tcpdump command with -A option. It will print ASCII text in recorded packets, that you can browse using page up/down keys.

tcpdump -A -r /tmp/http.log | less

However, tcpdump cannot decrypt information, so you cannot view information about HTTPS requests in it.

Clean up system space⚑

Clean package data⚑

There is a couple of things to do when we want to free space in a no-brainer way. First, we want to remove those deb packages that get cached every time we do apt-get install.

apt-get clean

Also, the system might keep packages that were downloaded as dependencies but are not needed anymore. We can get rid of them with

apt-get autoremove

Remove data of unpurged packages.

sudo apt-get purge $(dpkg -l | grep '^rc' | awk '{print $2}')

If we want things tidy, we must know that whenever we apt-get remove a package, the configuration will be kept in case we want to install it again. In most cases we want to use apt-get purge. To clean those configurations from removed packages, we can use

dpkg --list | grep "^rc" | cut -d " " -f 3 | xargs --no-run-if-empty sudo dpkg --purge

So far we have not uninstalled anything. If now we want to inspect what packages are consuming the most space, we can type

dpkg-query -Wf '${Installed-Size}\t${Package}\n' | sort -n

Clean snap data⚑

If you're using snap you can clean space by:

• Reduce the number of versions kept of a package with snap set system refresh.retain=2

• Remove the old versions with clean_snap.sh

#!/bin/bash
#Removes old revisions of snaps
#CLOSE ALL SNAPS BEFORE RUNNING THIS
set -eu
LANG=en_US.UTF-8 snap list --all | awk '/disabled/{print $1, $3}' |
while read snapname revision; do
    snap remove "$snapname" --revision="$revision"
done

Clean journalctl data⚑

• Check how much space it's using: journalctl --disk-usage
• Rotate the logs: journalctl --rotate

Then you have three ways to reduce the data:

1. Clear journal log older than X days: journalctl --vacuum-time=2d
2. Restrict logs to a certain size: journalctl --vacuum-size=100M
3. Restrict number of log files: journactl --vacuum-files=5.

The operations above will affect the logs you have right now, but it won't solve the problem in the future. To let journalctl know the space you want to use open the /etc/systemd/journald.conf file and set the SystemMaxUse to the amount you want (for example 1000M for a gigabyte). Once edited restart the service with sudo systemctl restart systemd-journald.

Clean up docker data⚑

To remove unused docker data you can run docker system prune -a. This will remove:

• All stopped containers
• All networks not used by at least one container
• All images without at least one container associated to them
• All build cache

Sometimes that's not enough, and your /var/lib/docker directory still weights more than it should. In those cases:

• Stop the docker service.
• Remove or move the data to another directory
• Start the docker service.

In order not to loose your persisted data, you need to configure your dockers to mount the data from a directory that's not within /var/lib/docker.

Set up docker logs rotation⚑

By default, the stdout and stderr of the container are written in a JSON file located in /var/lib/docker/containers/[container-id]/[container-id]-json.log. If you leave it unattended, it can take up a large amount of disk space.

If this JSON log file takes up a significant amount of the disk, we can purge it using the next command.

truncate -s 0 <logfile>

We could setup a cronjob to purge these JSON log files regularly. But for the long term, it would be better to setup log rotation. This can be done by adding the following values in /etc/docker/daemon.json.

{
  "log-driver": "json-file",
  "log-opts": {
    "max-size": "10m",
    "max-file": "10"
  }
}

Clean old kernels⚑

!!! warning "I don't recommend using this step, rely on apt-get autoremove, it' safer"

The full command is

dpkg -l linux-* | awk '/^ii/{ print $2}' | grep -v -e `uname -r | cut -f1,2 -d"-"` | grep -e [0-9] | grep -E "(image|headers)" | xargs sudo apt-get -y purge

To test what packages will it remove use:

dpkg -l linux-* | awk '/^ii/{ print $2}' | grep -v -e `uname -r | cut -f1,2 -d"-"` | grep -e [0-9] | grep -e "(image|headers)" | xargs sudo apt-get --dry-run remove

Remember that your running kernel can be obtained by uname -r.

Replace a string with sed recursively⚑

find . -type f -exec sed -i 's/foo/bar/g' {} +

Bypass client SSL certificate with cli tool⚑

Websites that require clients to authorize with an TLS certificate are difficult to interact with through command line tools that don't support this feature.

To solve it, we can use a transparent proxy that does the exchange for us.

• Export your certificate: If you have a p12 certificate, you first need to extract the key, crt and the ca from the certificate into the site.pem.

openssl pkcs12 -in certificate.p12 -out site.key.pem -nocerts -nodes # It asks for the p12 password
openssl pkcs12 -in certificate.p12 -out site.crt.pem -clcerts -nokeys
openssl pkcs12 -cacerts -nokeys -in certificate.p12 -out site-ca-cert.ca

cat site.key.pem site.crt.pem site-ca-cert.ca > site.pem

• Build the proxy ca: Then we merge the site and the client ca's into the site-ca-file.cert file:

openssl s_client -connect www.site.org:443 2>/dev/null  | openssl x509 -text > site-ca-file.cert
cat site-ca-cert.ca >> web-ca-file.cert

• Change your hosts file to redirect all requests to the proxy.

# vim /etc/hosts
[...]
0.0.0.0 www.site.org

• Run the proxy

docker run --rm \
    -v $(pwd):/certs/ \
    -p 3001:3001 \
    -it ghostunnel/ghostunnel \
    client \
    --listen 0.0.0.0:3001 \
    --target www.site.org:443 \
    --keystore /certs/site.pem \
    --cacert /certs/site-ca-file.cert \
    --unsafe-listen

• Run the command line tool using the http protocol on the port 3001:

wpscan  --url http://www.site.org:3001/ --disable-tls-checks

Remember to clean up your env afterwards.

Allocate space for a virtual filesystem⚑

Also useful to simulate big files

fallocate -l 20G /path/to/file

Identify what a string or file contains⚑

Identify anything. pyWhat easily lets you identify emails, IP addresses, and more. Feed it a .pcap file or some text and it'll tell you what it is.

Split a file into many with equal number of lines⚑

You could do something like this:

split -l 200000 filename

Which will create files each with 200000 lines named xaa, xab, xac, ...

Check if an rsync command has gone well⚑

Sometimes after you do an rsync between two directories of different devices (an usb and your hard drive for example), the sizes of the directories don't match. I've seen a difference of a 30% less on the destination. du, ncdu and and have a long story of reporting wrong sizes with advanced filesystems (ZFS, VxFS or compressing filesystems), these do a lot of things to reduce the disk usage (deduplication, compression, extents, files with holes...) which may lead to the difference in space.

To check if everything went alright run diff -r --brief source/ dest/, and check that there is no output.

List all process swap usage⚑

for file in /proc/*/status ; do awk '/VmSwap|Name/{printf $2 " " $3}END{ print ""}' $file; done | sort -k 2 -n -r | less