Ubuntu and Debian images built by Armbian got popular because of the sad state of affairs of most single board computers. But since Raspberry Pi boards are rather well-supported by the Raspberry Pi Foundation and community, plus part of it is closed-source, there’s been little motivation by the Armbian community to work on it, and that means there’s no official support for Raspberry Pi.
But some Armbian test images have just been released for Raspberry Pi 4, specifically Ubuntu 22.04 “Jammy Jellyfish” images with a CLI (headless) version, and two desktop variants with Cinnamon and XFCE desktop environments which you can download on their server(s). Separately, DietPi 7.9 lightweight Linux distribution based on Debian has also been released with various improvements.
The good news about Armbian images for Raspberry Pi is that they exist, but the less good news is that those are still “test images”, and the popular board is not yet an officially supported Armbian SBC. 9to5linux quickly tested Armbian nightly builds on Raspberry Pi 4, and found most features to work, except audio. There’s still no maintainer for the Raspberry Pi 4 images, and if you’d like to get involved you can apply to become the official maintainer for Raspberry Pi Armbian images. If nobody volunteers, support will be dropped from Armbian.
DietPi has been around at least since 2015 to provide lightweight Debian images for Raspberry Pi boards, but also other models such as ODROID, Radxa, and Pine64 single board computers. The latest DietPi 7.9 adds password protection for DietPi-Dashboard, multi-backup archives for DietPi-Backup, enhanced Apache PHP setup (php-fpm support), as well as better GPU accelerated video playback on Raspberry Pi, among other changes. DietPi supports Debian Buster, Stretch, and BullsEye Debian, but DietPi 7.9 will be the latest release for Debian 9 “Buster” that was first outed in 2017. A more detailed changelog can be found in the announcement.
Jean-Luc started CNX Software in 2010 as a part-time endeavor, before quitting his job as a software engineering manager, and starting to write daily news, and reviews full time later in 2011.
Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress
Armbian on a Raspberry Pi? Have the heretics taken over the cathedral?
Maybe the most funny part is that this happens at the same time those Armbian guys want to drop ‘support’ for ~50 boards since they’re too overwhelmed. It took them just years to eventually get that quantity destroys quality after adding SBC after SBC to the build system.
And then this 🙂
The other funny thing: what separated Armbian in the past from other ‘distros’ was low-level stuff and some love for details (kernel/driver stuff, tuning settings).
Here it’s quite the opposite: it’s just Ubuntu for RPi 4 (including ThreadX from RPi Trading Ltd. and fortunately also their wireless firmware package) garnished with some Armbian scripts that are partially unmaintained since years. No settings/tuning but instead ridiculous overvolting that fries the CPU cores all the time. Armbian finally went DietPi…
My impression always was that major aim of Armbian is (at some point became) to get all SBCs running with mainline kernel, which is the only way that makes it possible to support such a large number of SBCs without becoming overwhelmed. Means, when moreless unmodified upstream Linux runs on RPi 4, it shouldn’t cause that much extra work. But the devil may be in the details. I see the potential long term benefit that upstream Linux gets better support for Raspberry Pi in general, driven also by issues/requests sent by Armbian/users, and the other way round RPi Foundation may move features gradually from closed source GPU blob to open source upstream/standard solutions, like it was done with KMS and V4L2+libcamera based decoders recently.
Not sure which missing settings/tuning you mean, I know that Armbian has an SMP affinity optimisation script, though automated without user input/choice(?), but at least with official RPi kernel this is not possible (write-protected). DietPi scripts are heavily maintained with lots of settings, tuning, including overclocking, but also power savings and CPU+I/O scheduler settings etc. But DietPi does no own kernel development, so utilises the available firmware/boot/kernel parameters/settings which Raspberry Pi repository and for many SBCs also Armbian kernel/bootloader provides. So the limited development time goes into providing ready-to-run solutions for software/server stack installs and (userland) system config tools to end users. Aside of the kernel development, it has overlapping aims and tools with Armbian, but definitely different priorities, with a high focus on software install+setup solutions.
> major aim of Armbian
…never has been defined. The most important contributor to Armbian ever (Mikhail) tried to discuss this with ‘Armbian owner’ just like me over and over again but to no avail. He retired earlier than me.
Wrt comparison Armbian vs. DietPi the main point (at least for me) is: can you trust the OS image? The Armbian build system allows you to build your own image from scratch so you only have to trust Debian/Ubuntu ‘staff’ and those involved in Linux Kernel. In theory. In practice Igor always gets in your way and uses his role as ‘Armbian owner’ to introduce stability/security flaws as it pleases him. In the past he just bricked boards since he has no idea what ‘stable’ means, but recently he focuses on totally unnecessary security flaws on your device: https://github.com/armbian/build/commit/f0f10a5b68aff3c766f2e8790ac4ce9f3c3e2160
DietPi takes/took some random Debian ARM images found somewhere on the net, ran some scripts for ‘diet’ measures (which honestly majority is plain stupid or ‘as effective as diet recommendations for human beings‘) and then advertised this as an own ‘distro’.
While in reality it was just ‘something found somewhere on the internet’ garnished with some scripts. Remember when you DietPi guys realized some ‘nasty’ cron job on an OS image you pulled from FriendlyELEC? Similar thing was on a Pine64 image? Some crappy q&d hack to improve NAS performance? Users reported it and only then you noticed. But it could’ve been a hidden backdoor (or thousands) since if you grab something off the Internet to base your scripts on you simply have not the slightest idea what you serve your users as download.
Funny thing is: since Igor is not involved in package management for the RPi4 image it’s not possible that he bricks your RPi since… not involved into his usual u-boot/kernel update drama. As such the rpi4b image could be the one Armbian image that is most stable.
But… since this Armbian image also pulls in an Armbian repository… as an ordinary user you’re simply lost. If Igor decides to be in playground mode again he simply will introduce a new security flaw that gets pulled to your device the next time you run a ‘security’ update.
While scripts are not updated automatically for different SBCs on an automated build system, there’s a drawback compared to conventional x64?
Diversity will increase with new risc-v boards and manual (human) interaction is needed for adjusting all hardware for consumer OS.
Big companies will adjust this for (their?) customers and consumers can decide, what companies they like at publicly informed moment(s).
Thanks for the work for improvement for the people.
Let’s hear the wise folks for rpi5 OS enhancements and/or standardization?
Great civilizations know that supportive mood is support to health.
See this planet and know it’s now not then.
> tried to discuss this
Your opinion was and still is valuable, but if we couldn’t get on a common grounds … It takes a lot of persistency, waiting, organisation and perhaps we didn’t find that together in right order to make it happen. That’s life. I never deliberately blocked you, Michail or anyone. But you should and could easily block me if I was doing wrong. You always had this power. In fact you never lost it. You just decided to not help since we deliberately refuse your ideas(?), because we have different perspectives (which is actually a good thing) on what should or shouldn’t be done.
If person(s) gets derailed, and that is happening all the time, we deal with that. When possible. Sometimes takes days, sometimes months. Putting a blame on someone is easy. Just like demanding that bug is fixed and providing nothing in exchange. Yes, I accept my guilt. Now what?
Dealing with people is much harder than solving technical problems. That is not any secret. Any nobody want to deal with those problems. And we are always short with everything. I do put people / team / contributors / doer first, users second. Users wants their problem first. It’s a constant violent fight for that attention, so in such aggressive environment, which is a new experience for many, it’s hard to function. Not every person that would like to tinker technical matters with happiness will survive that.
Each group of people has internal fights going on. Constantly. I regret if someone leaves because things can’t be resolved in some common ground, perhaps it was not noticeable until it was too late. Sometimes things just doesn’t work out.
> since Igor is not involved in package management
I don’t even try to take responsibility over the whole release since that is not possible. Not even with “1000 people and best possible organisation”. Anyone else can pretend that, I won’t. But since nobody shows up to do things better, we have no choice but to roll things best we know. We have improved processes greatly. https://docs.armbian.com/Process_Release-Model/ They are still not perfect, but they are much better as you know them.
Also whole packaging was refactored, not by me, (I should avoid critical changes, since I am nothing but a troublemaker … ) but another nasty bug was introduced which was fixed at once when found, but some bugs needs to be find and will wait years to be fixed since users don’t care. You / users don’t care. We care.
You know how things are and you pin point to a bug I have created. Well, I have created more of them. Why? Because I am software developer. We are the one that create bugs.
This is open source and Linux is a work of thousands of people. People have no idea that this is community work. This perception is foreign to them. 90% expects we will jump up and fix a bug, found in random manner, at once. They don’t ask if the bug is ours and who will pay for resolving it. “It is our duty to provide fully functional software”. Or they will not donate 🙂 This always put smile on my face. Like their donations would have any impact on anything. They don’t realise we donate to them and not vice versa. Who is financing R&D, support and maintenance? They request 10.000 EUR services virtually every day and they are willing to pay 1 EUR. But only if we will do it at once and in a polite manner. Some things didn’t changed …
Thomas, you have all rights for your reality as we all do. I never denied mistakes and they are being corrected in best possible manner. Since this is the only possible way. Not just with Armbian, but generally with all distributions and community supported open source software at large.
Bugs in FOSS world are introduced daily and sometimes takes years to fix them. In professional world its not much different regarding this, just they are paid and most FOSS people are not. And you have little control over them by default. Perhaps that is the source of many troubles here?
If the process is wrong, you can always improve it. But you have to invest into the process remaking, not just asking to be changed. If wrong people are at wrong places, you provide replacement for their roles. Complaining that they are wrong without providing solution? Replacing one problem with another usually solve little to nothing.
But if you have no people and constant violence from users side, people go first because of that. Second also because goals are not unified, leadership of internal team is handled poor if anything at all. If this is lead by community needs, they would eat this team alive. And you know that very well. A lot of resources goes into protecting doers from consumers of their time.
Without making mistakes, if everything works, you don’t learn. And learning is the key important thing in the development process. We were not born smart and even we had decades of experiences in the industry, we still know little.
Here we go with 10.000 EUR per thing again. It become really pathetic to read this stand on every occasion! Why don’t you just shutdown the operation or at least the forum. And then there will be no more user requests and you could do whatever you like with the project (to be your playground). I though projects like this are oriented toward users. Seems not.
Yeah right. There are hundreds of abuse attempts on Github as well. On email, everywhere.
Shutting down forum won’t help.
It is. You are welcome to step up and serve them. We have no capacity:
https://forum.armbian.com/staffapplications/
There are too many with too expensive questions. Some are projects for 100.000 EUR (to not use your favorite number). But it doesn’t matter. They don’t even think money is required for development. Free software is developed for free. Especially software developed on demand, yeah right. We don’t even think to talk them down or do anything about. The problem is that they are so far away from reality that it hurts. And this is kind of hackers community, not some Facebook / Kali hackers linux wanna be world.
I think you are far away from reality.
I don’t use Armbian and I hope I never will. I don’t like sw with such toxic attitude and environment. That’s why I have no intention for any help. But I do free work almost daily somewhere else. That’s why I understand projects like this.
Everyone has its own reality and we can’t expect ours will match. Equally distant could be your reality from the “right” reality, don’t you agree?
You don’t help me but random people that happens to hang around our forums.
You are among those very few, that contribute. Helping people is nice and good things, but it has limits. Which is, in our circumstances, very easy to cross. You are perhaps not there yet and I hope you never will be. But if you will be one day, you will start to understand our pain. It is not simple to turn people away, when they need you most. But you have to triage. It’s your family, your mental health … or their problem without any compensation whatsover.
I know that. We also do. A lot. But people are not happy with that. They want more. And more. And more.
Assuming to understand project like this is wrong in fundament. To understand, you need to discuss about without sticking solid to existing standpoints.
Your have ofc rights to do that. I also don’t ask or expect that everyone would like me or the project we do. This is everyone personal choice.
Probably never been to Debian forums?
I respect all of my clients, people that contribute to open source, but this topic is not about them.
And who are the clients? The one who are quiet and just use Armbian and doesn’t need anything else (because everything they need works for them) or someone who do ask? My opinion is that you just don’t like this second group. Or maybe only who pay in advance.
And you usually saying how time is valuable: Why do you need to write such a long novels like in this thread. It is waste of time.
I have a few clients that helps paying the bill so we can have this party. Supporting random people from the internet is not included.
If you have downloaded software you also made a contract. You are from the FOSS community. You should know very well, what kind of contracts are between users and creators. You don’t get any client status by downloading software with GPL license.
When “clients” thinks that only they and their time matters, you know how you will deal with them. If you will choose to deal with them anyway, it’s your choice. We also do, but we would also like to live our lives, so they are turned down. Some by ignore, some with verbal denial, some with – go and hire someone (this is written in a support contract as well), some by force.
Answering blank is already a huge expense since each “customer” support exchange costs. And when you make no money from this interaction, this means you are losing it. When this is 100 EUR per month, who cares. I can easily afford to donate such amount to community. Even 1000 is not a problem. But at some point you have to stop giving away.
We are also asking them to help, at least a little, like I have asked you before, so we are able to help. But most again only look on their egoistic matter, ignore that message completely. They came here to be served, to get attention that only their personal problem is solved. And they don’t even think to provide nothing in return. Who is creating toxic environment? Doers or spoiled consumers? The same is in many places, on vendors forums, Armbian is nothing special about. At least our “clients” have no moral right to ask for anything.
Basic rule of any trade. Money is taken in advance when you don’t trust that person will pay for your services. Not necessarily with money. Money is just universal exchange.
I am happy to clarify all this, I am not so happy to be someone slave.
The point is in control over your personal time and satisfaction how it was spent. With who. I have to say no, to spent time with family, friend. Purpose matter most.
Reading all such discussions still gives me the impression that you don’t know what you want. Simple as that.
Igor, Thomas,
it’s very interesting for me to read your discussion, because I’m seeing some reciprocal misunderstanding that seems to stem from overlooking some parts that I’ve been dealing with for a now two decades without really choosing it, like most of us in the FOSS world.
In my situation, my activities evolved to the point where 100% of my work time is spent on it and a respectable company was built around the project, to still pay a tens or two of opensource developers to make the project continue to progress, to follow evolving standards and user needs, and to fix issues. As you can imagine, for a company to run like this, it needs a lot of money and it’s certainly not donations that make us eat at the end of the month (donations are only made to help donators materialize their happiness). It’s customers that make us leave, paid customers.
So one could think that our most important asset is our customers. But that’s wrong! Our most important asset is the project’s reputation! Because with a good reputation you easily find customers. With customers you don’t buy a reputation. And who gives us our reputation ? For the most part our community users!
That’s why we devote extreme amounts of time to deliver them quality and fix the issues they report. Some weeks we can spend up to 80% of our time on issue reports. That’s horrible, but the flood of “thank you”, “wow that was fast”, “amazing work” etc that flows back gives the project the reputation that indirectly makes our living.
Igor, my feeling along all your comments is that you think that you’re only giving to the users and that you get nothing in return, and by doing this you seem to be completely ignoring this extremely important aspect of reputation. A linux distro isn’t often discovered first via a google search. Many times it’s one person talking to a friend or coworker about how he wasted his whole week-end on a kernel bug or shitty installer, and often the other one says “ah you’re using that board, have you tried $favorite_distro, usually it works fine on such boards?”. If the first thing you respond to your complaining users is “you have nothing to expect from me, I’m not your slave even if you give me $1000”, the person asking for help will not keep a good memory of that experience at all and will never recommend your distro to anyone. Worse, those that are lost in trouble will only remember that their last contact with the distro was when it failed, and ruined their board.
Of course it’s unfair, and as you said you cannot stop users from asking for help because they’re in a situation of extreme stress (i.e. “I put a month-worth of money saving on this board and now it’s dead due to some jerk not being able to check a command’s return value”), this quickly makes conflicts raise. I’m seeing this as well, when our users face trouble, it’s their whole e-commerce site that’s down. One difference is that ours are mostly professional while yours are mostly amateurs so that changes the tone of the discussion, as amateurs tend to expect a lot more because they don’t see where the value is. But at least, together with them you create value. While for you these happy users are an asset (they MUST be one), for them they get your work for free so they give nothing, and the vast majority are always willing to be happy if they feel respected. Someone who gets something positive for free usually spreads the word.
Like it or not, that’s the price to pay to grow a reputation. If you only work for the reputation of course you’re doomed. If you do it for the fun, it’s a thankless effort. It’s fun at the beginning and becomes a full-time job with its lot of annoyances, to the point where the project can be very toxic to you, its creator (and I suspect from your tone that it reached that point). I’ve known such situations a long time ago and do not want to get back there. I suspect that’s where Thomas possibly underestimates how hard and painful that can be, because at some point there’s one person that has to decide whom to scream at if needed, and whom to protect from the abusers, and that’s never a pleasant job, it can quickly ruin your whole day at once, several times a week.
But if you can start to find paid customers, then you must completely change your way of seeing community users. You need to make them happy so that they spread the word that is needed for you to get more customers.
Making users happy is difficult. Instead, making them happier should be the main goal. For example instead of constantly trying to make them feel bad for demanding things they do not deserve, just let them know the number of pending questions before them and how far they are in the queue. It’s exactly the same but they don’t feel insulted and instead they realize that you’re doing your best. They can continue to search during this time and share some findings and sometimes even find the solution and proudly announce it, thus saving your time twice.
I know how difficult it is to keep a running machine stable, to the point that I never ever trust a major kernel upgrade to reboot to a working shell prompt. Thus I know it’s not your fault when there’s some breakage that impacts users and that was inherited from mis-applied patches after an LTS update and such issues. I’ve read some of Thomas’ comments about making things more reliable and avoiding cosmetic changes in stable branches. I generally agree with him on that but I also know that some pragmatism is also needed sometimes. My personal experience shows that bugs, errors and even developers mistakes are all easily forgiven when users are satisfied with the response, precisely because they know they get it for free and they most likely couldn’t figure it by themselves.
I’ve used Armbian a little bit. I think I might still have it on 2-3 SBCs that are not often powered up, and have much appreciated the installation instructions which are among the hardest info to find for most boards. I do really appreciate the amount of work behind the curtains to deliver this. However I must say that reading your song about “how much do you pay me for this” doesn’t make me feel confident to recommend it to new users anymore nor to deploy it on 24×7 machines. Look at your response to Sander below, it was definitely not welcome and caused some confusion at least, maybe even a bit of pain for him to remind you that he donated. You should really control this a little bit more.
Don’t get me wrong, I’m not trying to judge or anything. I have not contributed to your distro (just shared a few info on forums a long time ago), it helped me a few times and that’s all, so I’m among those only helping to spread a good reputation. I’m just finding it sad to see that I could be hesitating to use it more regularly or to recommend it because the project reached a step that’s difficult to transition from. I know it’s not only up to you to help with this (that would be totally unfair), but you need to think a bit differently about the value of your users, and will then need a huge amount of help from others to help them even more. It’s not an easy equation but it’s not by seeing contributors quit on strategic disagreements that you’ll get it fixed.
I feel your frustrations Thomas, especially with the high quality work you had contributed. Some of the decision making processes were quite frankly insane. I also seem to remember many people in the past being shot down in flames and treated like idiots for merely suggesting that the RPI 4 would be a good way to expand Armbian’s user base, contributors and supporters. The general attitude was that RPI users were too stupid to deal with back then
I love the Armbian project and it has many great people. Igor works very hard and goes above and beyond in terms of effort and there are lots of incredibly skilled and motivated people helping the project. Where it fell down for me was the inability to create a coherent strategy and stick to it. The failure to concentrate on what could be done with the resources available and to continually over reach caused severe degradation in the quality of “stable” images over time. It doesnt matter how great your project management system is if your board doesnt boot back up after a kernel update
I have also tried to help in the past in a much smaller way than you. I donated VM’s and bandwidth on Azure for distribution mirrors and test builds / CI but they barely got used. I tried to donate 3-4 months project manager time from staff at my company and she didnt get a response from the team. There was too much interest in chasing buzzwords than defining strategy and getting things done so in the end I walked away
≥ I have also tried to help in the past in a much smaller way than you. I donated VM’s and bandwidth on for distribution mirrors and test builds / CI but they barely got used. I tried to donate 3-4 months project manager time from staff at my company and she didnt get a response from the team. There was too much interest in chasing buzzwords than defining strategy and getting things done so in the end I walked away
Yep I remember. It was very generous, sadly we proved to not be ready at the time.
Where it fell down for me was the inability to create a coherent strategy and stick to it. The failure to concentrate on what could be done with the resources available and to continually over reach caused severe degradation in the quality of “stable” images over time. It doesnt matter how great your project management system is if your board doesnt boot back up after a kernel updateAbsolutely our biggest pain point. The design of the new “support” model is intended to improve the situation based on a few simple tenants:
* Reduce scope everywhere
* A named individual for each supported device that participates in the testing and release process
* Provide a consistent model for the different states of board “support” that is heirarchical and compatible with both vendors and individuals
Time will tell of course.
Due to very limited resources and fact that we pay for you and not the vice versa, speed of R&D is as is. I am pretty sure we have different views on what quality control is and how to achieve it. Old school dying technique is to stay in the past as long as possible and only tackle bugs. Other, more prestige and effective way is to develop QA team and techniques and stay in a bit more recent times. Especially for us, which we deal with new hardware, this is essential.
I am pushing those things, but hard work is what everyone avoids …
We had to adjust strategy from time to time, but to gain better quality, we won’t get it by limiting to one hardware. We could ofc focus just on Rpi, on sales and have fun, adding mirrors, record videos and sell BS to people that have no clue about Linux.
But then Armbian would not be Armbian. We would not have developers community. We would not be doing anything useful simply because Rpi world doesn’t need developers. They need consumers.
Quality control standards are bringing up from embedded world. We seek help, we have dedicated positions for that. Again, when / if people will show up, someone needs to deal with them. Lead them on day to day operations. Ofc this is again me, because nobody from the outside even see that this is needed.
Automated testing is standard in automotive grade Linux. In amateur Linux, this is not. And especially not in ship-and-forget single board computer support. Linux for hardware that are common considered as toys. Some are, some not.
We hear it about quality also from users, while they again do absolutely nothing to help. There are individuals who would like to test, but when you explain them this is a serious job, and that testing as they understand this task is not needed since we have automation to find bugs. We have no ability / resources to fix what is found.
I know very well how to get better quality, but we don’t have such circumstances.
My mistake was accepting a gift, because I should knew we will not be able to do anything. We didn’t have anyone (but me, overloaded with everything) dealing with infrastructure, so it should be you to take care.
I didn’t repeat this mistake. We have received many such kill-your-time gifts so I denied most of them. Why would we accept a free server if we have nobody that would maintain it and paying electricity bill was usually not included. Old machines sucks a lot of power and gives little bang. And we had nobody dedicated that would develop and maintain core infrastructure.
At that time, we were simply unable to process gifts that needed additional work.
Today we have a more or less dedicated team that is maintaining infrastructure. Process of adding a new mirror or runner to the system is now on the level you imagine it was 4-5 years ago. Professional & fast. Now it is: https://github.com/armbian/mirror
Only this year we managed to bring CI on satisfactory levels. It was again a project for several months of work. This is not something you download, click here and there and you are done. While our clients doesn’t give us time to breath.
https://github.com/armbian/build/actions
Currently we have around 300 CPU dedicated cores ready to deal with CI workflows. More than enough. Some donated, mainly ours. We also maintain this infrastructure on our own.
Why not before? Because development takes time.
I remember the same. Then our emails must have been lost somehow, but it doesn’t matter. We tried, it didn’t work out, not a problem. I never argue about the performance of or push to donated resource.
I am still grateful that there are people that want to help! We would not get this far without. Thank you.
> The general attitude was that RPI users were too stupid
Nope. At least not for me. It was about overwhelming support efforts.
Users that were trained to rely on crappy scripts will get angry (if there’s no user called ‘pi’ living in ‘/home/pi and those other ‘hard coded’ stupidities’ the RPi world is full of).
three pages, each expertise for difference in perspective
https://worldpopulationreview.com/country-rankings/average-iq-by-country
https://www-laenderdaten-info.translate.goog/iq-nach-laendern.php?_x_tr_sl=de&_x_tr_tl=en&_x_tr_hl=de
https://ourworldindata.org/intelligence
Conclusion? If we don’t hear peoples problems, we won’t find a suitable solution for all?
It depends a lot, where world wide users are from and what’s their possibilities, thus discussion on support gets more diverse view points.
It’s about education and that’s responsibility from societies more advanced providing possibilities (that’s not about forced, stupid assimilation, having power doesn’t mean being educated, AFAIL).
According to your comments on CNX. you seems to only use Raspberry Pi? 🙂 Perhaps you start to support it, because nobody around Armbian still has no interest to do that. This way you will get opportunity to understand what Armbian does today and change things on better. It might surprise you.
There is no need to be nostalgic about things which were absolutely correct but made wrong. We still do the same, just better, we all have learned a lot and we still do mistakes. Yes, we were always on the edge with a board count and focusing too much into their details, but we managed to hande largest cheap HW group – TV boxes – extremely well.
Standards went up. Also and mainly because of our work. Also because of your work and others. Vendors have cash – let them hire people and pay for better support or just stop their business.
Process of fixing mistakes is never ending game as this is not the only (known) mistake we made. Have you perhaps noticed how much cheap junk came to the market in past few years and how many Armbian has denied? Vendors are getting more and more aggressive and crooked.
Since not many help in common good and this will obviously never change, things still needs a lot of time and huge portion of persistency to move on. Project is growing all the time, but not because of large board count – that can deceive you easily.
We started to plan to change support rules and how to enforce them to not totally blow about 6-12 months ago. After a year of development we also added another major improvement to the daily maintained build framework which is used by many 3rd party »improved distributions«. We also added UEFI arm64 (by the way, for our internal needs, also x68) support and we have a good plan to maintain this better.
Rpi was accepted as a is and put out mainly to see how people will react on this. We waited to change rules first, so we could do it. Many people anyway doesn’t care about any rules, doesn’t read anything and they will try to abuse them … vendors as well. Rules are one of our defence weapons and they were upgraded. Fight goes on.
There are people who needs Rpi with Armbian so they will have opportunity to build it – or download (if there will be a real name behind). If you / users want better quality in exchange for nothing – it’s anyway users problem first.
Pretty sure you can’t overvolt (dangerously anyway) any of the rpi models. If you think 1,06V is any dangerous voltage for the latest pi models (like the rpi4) you obviously don’t know much about hardware. 1.06V for 28nm based chip is nothing. Yes, it’s hot (but that’s due to insufficient cooling when no heatsink is installed) but it’s far from killing the chip. You need a lot more voltage for that and as far as rpi’s are concerned, you can’t do that in software anyway.
You can though its not easy and not something likely to happen by accident. I was seeing how high I could push an overclock on the Pi4 and had 8 units to play with because the silicon lottery isnt always fair and you get good and bad SoC’s
With force_turbo and over_voltage=10, one of my units died completely. Yes I know that these settings voids the warranty but I wasnt bothered if a couple died, I just wanted to see how far I could push them
I think someone messed up an april fools joke somewhere 🙂
DietPi also deals with Native PCs and Virtual Machines (VMware, Hyper-V, VirtualBox). So you are able to have the same environment on total different platforms. A VMware system is setup on a PC within < 5 min, without SD card flashing etc., this speed can e.g. be used very good for quick tests.
Also running a lightweight VM with X11 to run your favourite tools on a Windows PC is a nice option.
As Canonical provides a Ubuntu image for Raspi4 (https://ubuntu.com/download/raspberry-pi) … what is the advantage of Armbian providing one too? Tweaking?
Armbian is a build system first. Second. It mixes best of embedded and server / desktop technology. We are coming from low level world avoiding what most of distributions out there do – provide different wallpaper / outlook to make a new distribution and focus on sales. We add magnitude more. https://docs.armbian.com/Release_Changelog/ Yes, we also create bugs because of that and some people have issues with that, but does nothing to change that. In case of Ubuntu, we add tweaks, yes, but also remove Canonical proprietary things by default. Since the day 1. All builds can be easily re-assembled on user level, also at CI level. We are bringing embedded standards below desktop Linux. And (some of), for us, non relevant changes from lists as such https://itsfoss.com/linux-mint-vs-ubuntu/ which makes Linux Mint (or any other) better option than Ubuntu. Or worse. Depending on who is looking, distro hoppers (not our target group) likes so much. While as we know Mint is assembled from Ubuntu packages (and Debian version exists), they don’t deal with low level things much and they mainly focus into the desktop, provide their own SW installer, … What is their advantage? Beside not providing snap (we also don’t) … In theory – we could actually change userland to Arch based. Or something more exotic or rebuild all packages on our own, introduce own packaging manager … Why?
OK, thanks for clarifcation
I hope the Armbian people will provide Ubuntu / Debian images for upcoming RISC-V SBCs. I trust Armbian much more than a random Chinese suppliers with their fork-and-forget Linux images … and their Chinese documentation … via Telegram. Da Horror!
I am long enough in this business and on the world to understand many dark aspects of open source development. Which you would probably need to look into before going into – “please give me more”, “I need more” …
tl;dr;
Trust and credibility is our main asset. Yes. But that doesn’t cover expenses and damages users create. Every single day. Supporting their endless consumer appetite, business needs or pleasure costs. Real money, while users covers virtually nothing, they don’t even try to help. Most of them demands! “Its free, so I came here for free service” attitude. (mos) Vendors and others that (ab)uses our product free of charge also tries to look away and count their profits. But consumer of 3rd party products (we didn’t sell you anything) are keep asking us to fix your problems on our expense? LOL And users are stunned when they are told that they need a commercial support contract (pay for services they are asking for) in case they are not satisfied with the terms of low violent best effort support. And in most cases users are not satisfied since their perception of reality is years to decades away. (Dunning–Kruger)
No, we have no interest to do a move into that direction. But if you finance, we will add our money, expertise and trust to hire people that will move there. Volunteers are not piling up …
Igor, FYI: a few times, I’ve made financial donations to Armbian. And I’m an open source developer myself, including user questions (and yes, ignoring & blocking nasty users). So I guess your comment is a bit misplaced towards me?
Most if not all in my comments is general observation. Not personal, but it might sound this way because of my imperfect English communication. I am not native and will never be. No matter how hard I try.
I am grateful for users donations, but the harsh reality is that basic critical kernel development (for all Linux distributions out there) can’t be covered from this. In fact we can only cover some basic infrastructure costs and not a single hour of development or support per year. We add the rest and business donators. Majority remains our donation. Nothing wrong, until we came to violence of support. Then joy falls apart.
We have a lot less users than Linux Mint for example, so our reach to assemble their numbers are far away. Our development and support is for users critical, theirs is optional. Art. Ux, …
So you have a great product and a great business model. But if only, it wasn’t for the damned customers.
You cannot fix that kind of thinking.