My Homelab – wtf happened

What I started with

What I started with wasn’t actually this QNAP 4-bay NAS but two old HP Proliant DL380 G5. Those were some noisy motherf*ckers. I found them dirt cheap somewhere on ebay and decided I “needed” some of these servers. When I’d lived with my parents, they would have hung me for the noise and the energy costs of these machines. But – I had just become “independent” in my own little one-room apartment and I decided that some humming servers would be the best fit for me to experiment and host some gameservers while studying computer science.

The two servers arrived and were installed by me right beneath my desk. I was a little “surprised” when it turned these beasts on and I had a howling tornado siren in my apartment. But thank god – after the self test the servers were just as loud as a landing helicopter and maybe just a little bit noticeable after I put my headphones on. ( I had some interesting discussions on the Teamspeak Server if i should use Voice Activation). So I started experimenting and using these devices 24/7. I started with Hyper-V and some Windows “Server” administration (yeah – I now know better). Initial steps in creating VMs and Replication were done.

The Expanse…

I was already infected with the Homelab-Virus and in the meantime I bought the already seen QNAP NAS – which I quickly grew out of. I planned to use it as iSCSI Target and using it as a space to put my files, while still running my HP servers. I looked on Ebay for interesting findings and quickly became aware that I wanted servers that weren’t as noisy as the HP DL380 G5.

I found a reasonable priced 24U Rittal Rack and filled it with some Chenbro Cases. These were outfitted with custom hardware and were a lot quieter and then my old HP servers. They quickly replaced them. I also experimented with pfSense and flashing a WatchGuard x750 to use this Firewall Distribution.

Lucky findings on Ebay

While always looking for interesting, quiet and not so energy hungry servers, I came across the HP N40L Microserver. This was, what I should have bought instead of the QNAP NAS. It would have fitted my requirements better than the kind of unflexible NAS. Please don’t get me wrong – Network Attached Storage Systems (whether from QNAP or f.e. Synology) are quite useful – but not if you want to learn, customize a system to your needs and fiddle around with it in inappropriate ways. These things are not the servers I needed them to be.

The HP N40L is small but efficient server gear. I should have bought two and let them run in a Cluster and everything would habe been fine.

But I stumbled upon a Storage Server with an odd naming. Knowing from my long nights on Ebay – this was a Supermicro Case – i bid on them and got a 3U SC843 Case with Mainboard, 32Gb of ECC DDR3 ram and two Intel Xeon X5650 dirt cheap for less than 100€ (some when in 2013) .

It has grown …

I started planning projects, Exchange Server, Active Directory, Offsite Backup, additional JBOD Case, Webservers, Reverse Proxy, Nextcloud. A network infrastructure started forming in my head. I didn’t check with reality and was still acquiring hardware. I was a little bit too euphoric and excessive, sold the 24U rack and upgraded to a full size 48U rack.

This would not been possible if I hadn’t made some deals with mit girlfriend. She would get a full size shoe cabinet while I was getting my rack.

Current state

The current state of the Rack is the following (from top to bottom)

  • 1U TP Link Smart Switch
  • 2U -Hyper-V Server
    • pfsense
    • several webservers
    • is on 24/7
  • IP KVM
  • 1U Supermicro Project Server for Puppet
  • 1U Supermicro Server for Software Development Projects
  • 1U Supermicro Server (will become my offsite Backup)
  • 2U Supermicro Server (will become my onsite Backup)
  • 3U Supermicro Server (mentioned above)
    • 3x 4TB, 8x 2TB WD Red
    • 4x 256 Samsung Evo SSD
    • 96Gb RAM
    • Main File Server
    • Is also used by my girlfriend for gaming
  • HP DL 160 G7 – No CPU / RAM / HDD – needs to be sold
  • 2nd 3U Supermicro Server (Project still open)
  • APC 750 USV
  • APC 1500 USV

In the back is an full height APC metered PDU

Project Outlook

I’m actually trying to reduce energy costs as a kWh costs me around 29 Cents. So I won’t expand the amount of servers anymore. Maybe I’ll start to replace the old Hardware inside some server cases with some newer generation stuff. Replacing the power supply units to the SQ Models to reduce the noise would be a good idea too- I might be at the noise level I started with.

So – in the end there are still some projects left. As always.

“The Benefits and Challenges of Open Data in the Food Industry”

The title, this disclaimer and the following article were written by ChatGPT, a large language model trained by OpenAI. ChatGPT is not a human, but a computer program that uses artificial intelligence to generate text based on inputs it receives. The views and opinions expressed in this article are not those of the developers of ChatGPT or OpenAI, and the information contained in this article should not be taken as fact. This article is for informational and entertainment purposes only.

Open data of recipe and product databases has the potential to bring many benefits to consumers, businesses, and the economy as a whole. However, there are also some challenges and drawbacks to consider.

One of the main advantages of open data is that it allows consumers to make more informed choices about the food they eat. For example, my grandmother always used to keep a handwritten recipe book with her favorite dishes, passed down from her mother. With open data, this kind of family knowledge could be shared more widely and easily, allowing people to access a wealth of information about the ingredients, nutritional value, and cultural significance of different foods.

Open data can also help businesses, particularly small-scale producers and farmers, to gain a competitive edge. By sharing information about their products, they can reach a wider audience and build stronger relationships with their customers. This can be especially important for those who are committed to producing healthy, sustainable, and locally-sourced foods. Supermarkets like EDEKA, real, and Rewe, as well as online platforms like Chefkoch, could all benefit from open data by making it easier for consumers to find and compare products, and by supporting small producers.

In addition, some people argue that open data could lead to the consolidation of power in the hands of a few large companies, who could use their access to data to dominate the market (Source: [Research Paper 1]). This could potentially stifle innovation and limit consumer choice.

On the other hand, a paper published by researchers at the University of Oxford argued that open data could actually increase competition and diversity in the food industry (Source: [Research Paper 2]). They pointed out that open data could enable new entrants, such as small-scale producers and cooperatives, to compete more effectively with larger companies. This could lead to more diverse and sustainable food systems, and ultimately benefit consumers and the environment.

Overall, the potential winnings and economic possibilities of open data of recipe and product databases are significant, but there are also some challenges to be addressed. With careful planning and attention to the needs of all stakeholders, this could be a growing market that benefits consumers, farmers, and businesses alike.

One example of an open data project in the food industry is OpenFoodFacts, which is a global, collaborative platform that allows users to share information about food products. The project was started in France in 2012, and has since grown to include over 2 million products from more than 150 countries.

OpenFoodFacts is based on the idea of crowd-sourced data, meaning that anyone can contribute information about the products they buy and consume. This allows the platform to provide a rich and diverse dataset that is constantly updated and improved.

The data on OpenFoodFacts includes information about ingredients, allergens, nutritional values, and other product characteristics. This can be very useful for consumers who want to make healthier and more sustainable choices, as well as for researchers who are studying the food industry.

In addition to the data itself, OpenFoodFacts also provides tools and resources that make it easy for users to access and use the information. For example, there is a mobile app that allows users to scan barcodes and get instant access to product information, as well as a website that provides detailed search and comparison functions.

Overall, the OpenFoodFacts project is a great example of how open data can be used to improve the transparency and sustainability of the food industry. By harnessing the power of crowd-sourcing and collaboration, the project is helping to create a more informed and empowered consumer base.

In conclusion, open data of recipe and product databases has the potential to bring many benefits to consumers, businesses, and the economy. It can help consumers to make more informed choices about the food they eat, and can support small-scale producers and farmers by providing them with a platform to share information about their products. Open data can also stimulate economic growth and innovation, and could lead to the development of a “foodtech” industry. Although there are some challenges to be addressed, such as concerns about privacy and security, the overall potential of open data in the food industry is positive.



The wonders of UTF-8

or: Why “ä” and “ä” isn’t the same..

Don’t these characters look the same to you?
To me – they do. Well now they do – i noticed during a project that one of the characters didn’t show up on screen while being clearly visible in the Code Inspection Tools of Chrome or Firefox.

What had happend?
A colleague copy pasted text from a PDF File and used parts from it in a description text.

It seems that some software instead of using the simple “ä” use a UTF-8 combination equivalent of “a” and ” ¨ “.
Often the single ” ¨  ” is not contained in public available fonts. This character is called trema or dieresis.

Fortunately the php-intl package already contains a solution for my problem – the Normalizer Class: https://www.php.net/manual/en/normalizer.normalize.php

I attached an example for you:

$a ='ä';
$b ='ä';

echo urlencode($a);
echo ' ';
echo urlencode($b).PHP_EOL.PHP_EOL;

$a = Normalizer::normalize( $a, Normalizer::FORM_C );
$b = Normalizer::normalize( $b, Normalizer::FORM_C );

echo urlencode($a);
echo ' ';
echo urlencode($b);