Several of years ago I finally decided to invest time and money and build a home NAS. I had a bunch of photos from the DSLR scattered across numerous USB hard drives and decades worth of different "backups" so I wanted to centralize them. I did not want to vendor-lock myself and pay for the Google or Microsoft storage plans or similar cloud backup solutions and besides that, I did not want to family photos to their servers. So the only option remaining was to build or buy a NAS server. As I'm a web developer and ocassionally like to play with ops and tinker with linux servers, I decided to buy a cheap PC with good CPU, enough of RAM and throw in some WD RED drives. Although i have 2 x 4tb WD RED drives paired in RAID1 I still wanted to have some additional layer of redundancy.

So I bought another PC....

Few months ago, I moved into my own office. And what better place to put the backup server in than that?

Now I was facing only 2 problems:

1. servers being on different networks
2. which backup software to use

Servers being on different networks

This is the problem because servers can not communicate directly when they are on behind different routers, behind a NAT and/or don't' have a static IP.

TAILSCALE TO THE RESCUE!

Tailscale is kind of VPN but instead of on device acting as a central server, all devices registered are part of a network mesh. All devices will get a public IP so any device can connect to it over the internet without any need for port forwarding. Also, all devices can be pronounced as "exit nodes" towards the internet. By installing and setting up the tailscale on both servers, I was now able to make them "see eachother".

Which backup software to use

This was a tough one. I wanted to have incremental backups so I can backtrack files to some point in time but also in order to minimize the overhead space used. I also did not want to have some special backup file format from which I would have to restore the files, but rather just a plain and simple directory and file structure which I would be able to access from the backup server anytime without need for additional software or process.

At first I tried to set up urbackup.com for several days, but I struggled to do so for a number of reasons both on server (office backup server) and on client (home NAS server in this case) side. I looked at couple of different tools but none of them fit the descrption. I also tried to look for some existing rsync scripts that would fit my needs but didn't trust any of them. So I turned to ChatGPT...

rsync TO THE RESCUE!
I asked ChatGPT the following prompt:

I have two linux servers. One at home which serves as home NAS and the other in the office, which should serve as a backup server for some files of the first one. Both servers are using tailscale. I want to have scheduled remote backups for some folders of the home server to the office server. I want to keep monthly backups, one for each of 12 months in a year, i want to keep weekly backup, one for each week in a month, where last weekly backup should become monthly backup for that month, and I also want to keep daily backup for each day in a week, where last backup of the week will also become weekly backup for that week. I dont want to keep full backups. I want to have one full backup and then only add differences to subsequent backups. I dont want files on the backup to be in some weird fileformat for which i would need some software to restore. I just want them to be plain old files and dirrectories organized in some logical manner so I can simply go in to a daily backup and access the whole file list as if it was a full backup.

And this is what it's answer was:

You can achieve this setup using rsync with hard links to efficiently create incremental backups while keeping a fully browsable directory structure.

After some trial and error we were able to make up with the following process:

Step 1: install rsync on both servers:
sudo apt update && sudo apt install rsync

Step 2: Set up SSH access

1. Generate SSH keys on your home NAS:
   ssh-keygen -t ed25519
2. Copy the public key to the backup server:
   ssh-copy-id user@office-server-tailscale-ip
3. Test passwordless SSH:
   ssh user@office-server-tailscale-ip

Step 3. Create the Backup Script:

We'll use rsync with --link-dest to create incremental backups while making them appear as full backups.

#!/bin/bash

#Configuration - You will need to change the values below to suit your needs
SOURCE_DIR="/path/to/home/nas/folders/"  # Directory to backup
BACKUP_BASE="/path/to/office/server/backups"  # Base backup location on the office server
SSH_USER="home_server_user"
REMOTE_HOST="office-server-tailscale-ip"
BACKUP_USER="backup_server_user"  # User who should own the backups on the remote server

# Get current date components
DAY=$(date +%A)           # Monday, Tuesday, ..., Sunday
DATE=$(date +%d-%m-%Y)    # Format: dd-mm-yyyy
WEEK=$(date +%U)          # Week number (00-53)
MONTH=$(date +%m)         # Month number (01-12)
YEAR=$(date +%Y)          # Year

# Define backup directories with date appended
DAILY_BACKUP="$BACKUP_BASE/daily/${DAY}_${DATE}"
WEEKLY_BACKUP="$BACKUP_BASE/weekly/week_${WEEK}_${DATE}"
MONTHLY_BACKUP="$BACKUP_BASE/monthly/${YEAR}-${MONTH}_${DATE}"

# Find the previous daily backup (yesterday’s backup)
PREVIOUS_BACKUP=$(ssh "$SSH_USER@$REMOTE_HOST" "ls -d $BACKUP_BASE/daily/* 2>/dev/null | tail -n 1")

# Rsync options
RSYNC_OPTS="-a --delete --link-dest=$PREVIOUS_BACKUP"

# Perform daily backup
rsync -e ssh $RSYNC_OPTS "$SOURCE_DIR" "$SSH_USER@$REMOTE_HOST:$DAILY_BACKUP"

# Set correct permissions and ownership on the backup server
ssh "$SSH_USER@$REMOTE_HOST" << EOF    
    chown -R $BACKUP_USER:$BACKUP_USER "$DAILY_BACKUP"    
    chmod -R u+rw,g+r,o+r "$DAILY_BACKUP"
EOF

# Every Sunday, create a weekly backup from the last daily backup

if [ "$DAY" == "Sunday" ]; then
    ssh "$SSH_USER@$REMOTE_HOST" "cp -al $DAILY_BACKUP $WEEKLY_BACKUP"     ssh "$SSH_USER@$REMOTE_HOST" "chown -R $BACKUP_USER:$BACKUP_USER $WEEKLY_BACKUP && chmod -R u+rw,g+r,o+r $WEEKLY_BACKUP"
fi

# Check if today is the last day of the month
TOMORROW=$(date -d "tomorrow" +%d)
if [ "$TOMORROW" == "01" ]; then    
    ssh "$SSH_USER@$REMOTE_HOST" "cp -al $DAILY_BACKUP $MONTHLY_BACKUP"     ssh "$SSH_USER@$REMOTE_HOST" "chown -R $BACKUP_USER:$BACKUP_USER $MONTHLY_BACKUP && chmod -R u+rw,g+r,o+r $MONTHLY_BACKUP"
fi

# Clean up old backups (keeping retention policy)
ssh "$SSH_USER@$REMOTE_HOST" << EOF    
    # Keep only the last 7 daily backups    
    cd "$BACKUP_BASE/daily" && ls -t | tail -n +8 | xargs -I {} rm -rf {}  
    # Keep only the last 5 weekly backups
    cd "$BACKUP_BASE/weekly" && ls -t | tail -n +6 | xargs -I {} rm -rf {}
    # Keep only the last 12 monthly backups    
    cd "$BACKUP_BASE/monthly" && ls -t | tail -n +13 | xargs -I {} rm -rf {}
EOF

Step 4. Create the destination directory structure on backup server

mkdir -p /path/to/office/server/backups/daily
mkdir -p /path/to/office/server/backups/weekly
mkdir -p /path/to/office/server/backups/monthly

Step 5: Automate with Cron

crontab -e on home server

add: 0 2 * * * /path/to/backup_script.sh

This will:

• Run the script every day at 2 AM
• Log output to /var/log/backup.log (useful for debugging)

Step 6: Verify Everything Works

Test the script by running on home server:

bash /path/to/backup_script.sh

Then verify on the office backup server:

ls -l /path/to/office/server/backups/daily

Example directory structure you should end up with

/backups/
├── daily/
│   ├── Monday_27-03-2025
│   ├── Tuesday_28-03-2025
│   ├── Wednesday_29-03-2025
│   ├── Thursday_30-03-2025
│   ├── Friday_31-03-2025  <-- Last daily backup of the month
│
├── weekly/
│   ├── week_10_10-03-2025
│   ├── week_11_17-03-2025
│   ├── week_12_24-03-2025
│
├── monthly/
│   ├── 2025-01_31-01-2025  <-- Last daily backup of January
│   ├── 2025-02_29-02-2025  <-- Last daily backup of February (leap year)
│   ├── 2025-03_31-03-2025  <-- Last daily backup of March

Final thoughts:
✅ rsync's --delete option will delete all the files on the backup server that are missing or have been deleted on the home server. You might want to remove this one, it's up to you
✅ Ensures first backup is always a full backup (when no previous backup exists).
✅ Uses incremental backups (--link-dest) only when a previous backup exists.
✅ Automatically creates missing backup directories (mkdir -p $DAILY_BACKUP).
✅ Prevents rsync from failing due to missing previous backup folders.
✅ Maintains correct ownership (chown user:user) and permissions (chmod -R u+rw,g+r,o+r).

One of the first things I install when setting up a new Linux installation is htop - it's a defacto process manager for Linux command line. It has been around for quite some time, since 2004 to be exact. Recently, a modern, good looking replacement showed up - Bashtop.

Why Bashtop?

First of all, htop is written in C programming language which means it has some dependencies needed  to run graphical-like user interface in terminal. On the other hand, bashtop is entirely written as a single bash script so all you need to run it is bash shell and couple more utils which are pretty much standard to nowadays Linux distributions - awk, sed, GNU coreutils...

Fun fact - the main motivation for this app to be written in bash script is that the developer wanted to learn bash scripting

It provides much more information about hardware than htop CPU model, per-thread usage histogram, usage percentage and temperature. It also displays info about hard drives, network transfers etc. Also, it is super-easy to install and use, it looks amazing and refreshing.

It also supports themes! It comes with both dark and light themes out of the box. FTW!

These are some key features as listed on it's GitHub repo page:

• Easy to use, with a game inspired menu system.
• Fast and "mostly" responsive UI with UP, DOWN keys process selection.
• Function for showing detailed stats for selected process.
• Ability to filter processes.
• Easy switching between sorting options.
• Send SIGTERM, SIGKILL, SIGINT to selected process.
• UI menu for changing all config file options.
• Auto scaling graph for network usage.
• Shows message in menu if new version is available
• Shows current read and write speeds for disks
• Multiple data collection methods which can be switched if running on Linux

How to install Bashtop on Ubuntu 20.04 or Elementary OS 5.1

There are basically two methods and both are really simple to follow along. Whole bashtop app is a single bash shell script, a single executable file. No need to build or install anything.

Method 1: Manual installation using Git - Any Linux distro, OSX, FreeBSD

You will need to have Git installed before you continue.

cd ~
git clone git@github.com:aristocratos/bashtop.git
cd bashtop
make install

That's it! now you can close and reopen you terminal or simply reload your .bashrc or .zshrc and you are ready to go!

In your terminal run bashtop whenever you wish to have a look or kill some process.

If you ever wish to uninstall bash top that was installed this way, you can easily do that by running this command

cd ~/bashtop
make uninstall
cd ..
rm -rf bashtop

Method 2: Ubuntu and Elementary OS - install from ppa repo

Run the following commands in terminal

sudo add-apt-repository ppa:bashtop-monitor/bashtop
sudo apt update
sudo apt install bashtop

Uninstall as any other app installed using apt-get

 sudo apt-get --purge remove bashtop
 sudo apt-get autoremove

CPU Core temperature:

for ubuntu, elementary os, and other debian-based distros

One of the most awesome features bashtop provides is temperature for each individual CPU core/thread. For bashtop to be able to show the temperature, you will need to install one of the libraries that are can obtain and provide that info. Simply run this command in terminal and next time you run bashtop it will show you temperature readings for CPU cores.

sudo apt-get install lm-sensors 

You can learn more about installation on different linux distros on the bashtop GitHub repo

Options menu

Bashtop comes with various settings - from themes to refresh rate at which data on the screen will be refresh at. To access this menu just press ESC.

Select OPTIONS then press Enter key

Themes

There are 10 additional themes that come together with the installation. Select Themes option and press Enter key to tell bashtop to look for new themes in the themes directory.

 Change the theme using left and right arrow keys

Keyboard shortcuts

To learn about different keyboard shortcuts you can use, go to the OPTIONS menu, and then select HELP

Conclusion

Bashtop is certainly a refresh and interesting how terminal apps can look amazing.  I hope this article has inspired you to give bashtop a chance and you enjoy using it.

References and useful resources

https://github.com/aristocratos/bashtop
https://computingforgeeks.com/bashtop-resource-monitor-for-linux-macos-freebsd/
https://www.youtube.com/watch?v=nyx9AqaqUgA
https://www.youtube.com/watch?v=A6fdSHtPHAM

If you by any chance had an opportunity to use a trackpad on MacBook or even better a Magic Trackpad I highly doubt you weren't amazed. I know I was. It's definitly one of few things I missed the most when switched from MacBook Pro to ThinkPad T490. Eventhough ThinkPad's keyboard is by far superior to Mac's, the trackpad is... meh. Nontheless, it is a multitouch device and that means I can use it with swipe and pinch gestures to supercharge my workflow.

Sadly, I don't know about any Linux distro that comes with a usable gesture support. No worries, that's why this tutorial exists.

libinput is a library that adds various gestures support to the OS.

In order to be allowed to use gestures, your user must be added to the input group

In terminal, type the following command and enter your password

sudo gpasswd -a $USER input

After that, you will have to logout, and log back in order for the changes to take effect.

Install packages needed to install and use app for creating and mapping gestures.

sudo apt-get install wmctrl python3 python3-setuptools xdotool python3-gi libinput-tools python-gobject

After that, clone the libinput-gestures to your Downloads directory:

cd ~/Downloads
git clone https://github.com/bulletmark/libinput-gestures.git
cd libinput-gestures
sudo make install

To be sure that this utility is started whenever your computer is turned on, we need to run the following commands:

libinput-gestures-setup autostart
libinput-gestures-setup start

At this point, you can create a config file to map gestures to actions which I'll cover in another tutorial. For the sake of simplicity, we'll stick with a GUI app for this one.

To install Gestures GUI app, run following commands:

cd ~/Downloads
git clone https://gitlab.com/cunidev/gestures.git
cd gestures
sudo python3 setup.py  install

Mapping gestures to actions

Run Gestures app from the Applications menu.

Start by adding new gesture. Simply click + icon in the upper-left corner.

In the pop-up choose which type of gesture you wish to add - Swipe or Pinch. Next, choose a direction: Up, Down, Left, Right for the Swipe type of gesture and In, Out, Clockwise, Anticlockwise for the Pinch. After that, select how  many fingers the gesture is consisted of. In the Command field you should enter a command which will be executed when geture occures. For this, we will use xdotool which we installed at the beginning. It is a simple utility that programatically "fakes" keyboard and mouse events. Will use it to trick our OS to think we pressed a keyboard shortcut.

The way I use setup my gestures is:

4-finger swipe left perform xdotool key super+Right (Switch to workspace right)
4-finger swipe right perform xdotool key super+Left (Switch to workspace left)
4-finger swipe up perform xdotool key super+Up (
Toggle maximized)
‌4-finger swipe down perform xdotool key super+Down (‌
‌Multitasking view, same as super+S)
3-finger swipe left perform xdotool key alt+Left (
Browser back)
3-finger swipe right perform xdotool key alt+Right (
Browser forward)

Add getures one by one and click confirm.
For the swipe action, Two fingers option might be disabled because it's already provided by the touchpad itself for scrolling.  

As a last step, we can clean up  by safely removing cloned repositories

rm -rf ~/Downloads/libinput-gestures
rm -rf ~/Downloads/gestures

I hope this helps you in your everyday work, I sure know it helped me :)

Here are some links and references related to the libraries mentioned in this article

• https://github.com/jordansissel/xdotool
• https://www.semicomplete.com/projects/xdotool/
• https://github.com/bulletmark/libinput-gestures
• https://gitlab.com/cunidev/gestures/-/tree/devel
• https://www.youtube.com/watch?v=ArBCfhVsTZw

My name is Srđan Marjanović. I live in Novi Sad, Serbia. My list of things that I'm passionate about is ever evolving but some of the stuff that are always present on the list are of course programming, hacking and tinkering with computers and electronics, but also cooking, coffee, skiing, rollerblading, ice skating and photography.

I have almost a decade of professional experience as a web developer. Currently, my day job is being a full stack developer (with a sprinkle of Ops) @ Active Collab. I'm a part of one of the multidisciplinary scrum teams. My team is doing the magic behind the scene. We are in charge of servers, services and payments run smoothly. Also, we take care of our back office system that helps our support and success teams help our customers and our business development and design teams to better understand the needs of our clients. Every day we juggle with multiple services written using several technology stacks that are running both on AWS and bare metal servers. Some are real-time, all are highly-available.

Why am I doing this?

One of the main motives is that I always felt that I have many things to share with the world (outside of my soical media postings), but not just that. One way communication sucks. I would really want to spark discussions with all of you who come across this corner of the Internet.

On the other hand, I needed a place to articulate my thoughts and ideas and to have a reference of how I overcame some issues.

They say that if you really want to be sure you understand something, you need to be able to explain it to someone else, and that is my main motive - to learn by teaching.

Also, you probably noticed that I'm not a native english speaker and I hope that my english will improve while writing these posts.

What can you expect to find here?

As I'm passionate about some totally different stuff, it might get messy, but I will try to keep things down to:

1. Programming - PHP, MySQL, Golang, Python, JS, CSS, shell, refactoring, testing ...
2. Electronics - Arduino, Raspberry Pi, sensors, motors ...
3. DevOps - Jenkings, GitLab, Grafana, Prometheus, CI/CD, linux, servers, monitoring, automation ...
4. Productivity - routines, habbits, lifehacks, tools, methodologies ...
5. Personal - thoughts about books, videos, talks, shows ...

How often will I publish new posts?

It really depends. I have almost no free time at all as I'm working full time and I have several small kids to raise. I will probobly write in bursts in moments of inspiration and free time.

Until the next reading, have a great time