See the email thread on the misc list for more details.
In light of shrinking funding, we do need to look for a source to cover project expenses. If need be the OpenBSD Foundation can be involved in receiving donations to cover project electrical costs.
But the fact is right now, OpenBSD will shut down if we do not have the funding to keep the lights on.
If you or a company you know are able to assist us, it would be greatly appreciated, but right now we are looking at a significant funding shortfall for the upcoming year – Meaning the project won’t be
able to cover 20 thousand dollars in electrical expenses before being able to use money for other things. That sort of situation is not sustainable.
The OpenBSD project is the incubator for a number of other projects including OpenSSH and OpenSMTPD. If you use these or just want the project to survive, consider making a donation.
Again?
I thought we went through all this a few years ago, this time i have to say that i am sad that they have not figured out a better way of funding the project over the years than their current (same as former?) model. But if shutting down is what they have to do unless donations roll in once again it would be understandable. Perhaps they could be able to merge with a better funded BSD project somehow if all else fail.
Guess they didn’t sell enough t-shirts
On a personal style note, their artwork is fairly juvenile, imho. Not cool, not classy, more geeky than nerdish. This is why I never bought anything from them when I was seriously using openBSD ( although I did donate).
Money does have a funny way of disappearing like that – getting spent and all.
You have to be super rich if you want people to give you more money for having money: http://www.youtube.com/watch?v=9gE_oV80L1s
While I don’t use OpenBSD directly, I am concerned about this news. Many of us, including myself, use OpenSSH on a regular basis.
http://www.openbsdfoundation.org/donations.html
http://openbsd.org/donations.html
It’s not much, but I just gave 20 bucks. I’ve downloaded OpenBSD releases dozens of times, and felt that I should support a platform that has supported the rest of us.
While I applaud the openBSD folks for creating a rock solid OS, I have to wonder about their funding. With the recent revelations of the NSA I have to ask is anything secure?
Well, if you use the NSA’d encryption, then, no, OpenSSH is as weak as the NSA’d encryption.
If the NSA had filled it full of backdoors, surely they’d be worried about it running out of cash? TBH i’m more sceptical about the massive influx of cash into FreeBSD when that ran a “donation drive”. I’m sure if the NSA was backdooring any of these open-source projects, they’ll have a vested interest in keeping them active and well funded
Unfortunately, almost all billionaires were not geeks, so they will not sponsor a project like OpenBSD. It’s time for Bill Gates, Zuckerberg and Elon Musk to team up and support this project. .
I can imagine one billionaire will sponsor the project and create an awesome desktop operating system out of OpenBSD and then release it as Free software.
That would be pretty ironic, OpenBSD is a BSD not GNU-based system (like Linux).
GNU-based systems are considered ‘free software’, BSD is not. BSD is open source I guess.
No, BSD or rather the BSD license is considered a Free Software licence.
Quoting https://www.gnu.org/licenses/license-list.en.html “This is the original BSD license, modified by removal of the advertising clause. It is a lax, permissive non-copyleft free software license, compatible with the GNU GPL.
Maybe you were thinking about the CopyLeft vs Non-CopyLeft distinction between Free Software licenses
The Free Software Foundation has no problem calling the OpenBSD people fine people, I’m not so sure the OpenBSD people feel the same way.
The animosity between the two goes back to when Richard Stallman made a bunch of false accusations about the OpenBSD project including closed source software in their releases. The OpenBSD project had a right to be a bit upset about the sitation when it happened.
Some people dont understand that if you try and be a pure open source project, you can only rely on what few little things you sell. This includes CD sets, t-shirts and also donations. Thats the only way you can keep running without direct funding by any type of company or corporate entity. Some people see “sponsor” and they think huge amounts of money changes hands and thats not the case in all scenarios. Running a project like this myself, and I wish them the best of luck.
Roberto J. Dohnert
Head Developer
Black Lab Linux Project
http://www.blacklablinux.org
robertojdohnert,
Yea, I stopped maintaining my own linux distro a while ago. Each dependency was built from source using an automated build process. When I needed a new program, I’d add it to the build. I learned a tremendous amount in doing it, but it consumed alot of time and I wasn’t getting paid to do it. I would have continued it as a hobby, but after having a second child I lost the free time I used to have.
Being able to maintain your own OS impresses lots of people, but when it really comes down to it, it’s quite difficult to find a market for that kind of work.
What’s the name of your distro? I am just curious. Anyway, I think you can put it in you resume to impress potential employers.
allanregistos,
“What’s the name of your distro? I am just curious.”
There are no links to it now but it was gmlinux. It’s still running a few internal servers. I’m sad to let it go, but I wasn’t able to justify the effort.
Anyways there was a simple command line installer with a mere handful of settings involving which partitions to use. It worked on flash drives, hard disks and cdroms. It had three installation modes:
1. Normal file system
2. AUFS overlay on persistent storage
3. AUFS overlay in ram/swap.
Note that one could always mount additional persistent storage, the modes above relate to the OS itself, not the user data.
The reason for the overlay was to make it trivial to reset OS to a known state. The full OS was in an overlayed squashfs file system. If the machine got clobbered, fixing it was a simple matter of rebooting it. I was working on automatic app provisioning after boot with local caching to minimize traffic, this was a work in progress.
My basic idea behind gmlinux was that it would be installed on servers in a completely generic fashion such that they would take on application personalities from a pre-configured application repository. This keeps server maintenance to a bare minimum.
Although I still am fond of my original approach, I’ve migrated to stock debian running services inside of KVM instances that are isolated and very easy to customize as needed.
Your approach sounds a lot like Tiny Core. Very interesting both for those who know what they’re doing and for those who don’t have any clue..!
Or, another favourite: Porteus!
pepa,
I’d never heard of Porteus before.
TinyCore is very cool. Having applications follow you around between computers is awesome. I really liked that about Sun’s Solaris desktops at university. I haven’t tried any new distros in a long time, but now that you’ve brought it up I think TinyCore should be on my short list.
In my case though I was targeting headless servers exclusively, so there was no desktop at all. There were grandiose visions of running it in a massive data center I had to settle with running it on my amateur cluster.
MicroCore is tiny, headless (I mean, no graphical environment) and deliciously capable!
Anyway, aufs is pretty awesome..!
I have a suggestion. At least for me, it can be disappointing to hear a distro listing several DEs or supports several desktop environments at once. This is just one example of wasting resources on the part of the distro developer. I think a developer needs only to focus on one desktop environment and promotes it so there is no wasted energy on other DEs.
As a user, it is not really the desktop that matters most, what matters is the application availability and development tools. We need it to work on something is very important than we need a beautiful KDE desktop.
In my specific case, I need to install a Linux distro in a business environment. This should be the focus of Linux desktops, business, optimize your Linux desktop for business use, so what you use in your business, will naturally will be used in your homes too. (See Microsoft Windows). If your Linux desktop is not usable in business, then sorry, I cannot promote your Linux desktop, that is why, still, I am using Ubuntu in spite of the problems.
Another specific business case, Microsoft Office is really needed in an office environment, so a distro that installs Microsoft Office seamlessly with the aid of Wine with no Internet hoping reading tutorials here and there really makes sense than supporting all the Unix Desktop environments on the planet…
Regards,
Allan
We only support 2 Desktop environments, XFCE and KDE. On our Enterprise focused distro we only roll KDE now.
Roberto J. Dohnert
Head Developer
Black Lab Linux
http://www.blacklablinux.org
@robertojdohnert – Never heard of your distribution. Just checked it out and it looks great and am downloading it now.
BTW – Just bought an OpenBSD t-shirt to support the project. I don’t use OpenBSD anymore but still use OpenSSH daily. Thanks Theo and team!
I do sympathize how difficult it can be to contribute so much to the community while on the brink of financial collapse. I’d be sad to see OpenBSD fall. Yet I do have to wonder about what was said:
For a rough conversion…
consumer rates ~ $0.15/kWh
commercial rates ~ $0.06-$0.10/kWh
$20000/yr * (1kWh/$0.15) * (1yr/8760h) = 15.2kW
$20000/yr * (1kWh/$0.06) * (1yr/8760h) = 38.0kW
A power hungry 8 core + 6 disk dell server consumes ~ 300W full tilt. Apparently servers average ~ 500W including cooling:
http://www.vertatique.com/average-power-use-server
So they could run a loaded rack or two of build servers non-stop for $20k/year, but it really begs the question, does this project NEED that? Do they need to run so many servers all the time? Why not consolidate them using virtualization? I’d be very interested if anyone could shed light on it.
Maybe the usage could be optimized, even something simple like rebuilding the entire code base less frequently. Someone let me know what I am missing!
Because they support multiple architectures which can not be virtualised at all?
terra,
“Because they support multiple architectures which can not be virtualised at all?”
Kudos for thinking of that, I didn’t. They probably use a cross compiler, but they might have other servers for actual testing.
Still, qemu might be an option since it emulates many kinds of platforms, though I’m not sure how well.
If this really is a source of the problem, then perhaps it’s time to retire some of the archaic architectures they’re building binaries for? Just distribute the build environment so that users can compile their own binaries easily.
Hmm, in this email Theo de Raadt seems to be strictly against cost cutting solutions.
http://marc.info/?l=openbsd-tech&m=138973312304511&w=2
The email he replied to seemed well founded to me, especially if things are getting bad enough to threaten the whole project. It’s nice to have everything, but they need to focus on the critical things that keep them viable. Hopefully this is realized before it’s too late.
He can stay in denial all he wants but in the end if the funding doesn’t come forward they will be forced to do it anyway.
I was going to donate to the project until I saw that e-mail. He seems to think the world owes him money to keep development going and is completely against any practical solutions such as trimming architectures, cross-compiling, moving servers or virtualization. It is one thing to ask for one-time donations to help the project through a rough patch. It is entirely something else to throw money down a well knowing the project managers aren’t going to make necessary adjustments.
He actually justifies it very well, he says that he would probably lose half of the current developers are the ones that like devving for these systems. The last time I checked it was 80 developers.
Which makes more sense, potentially losing 80 developers for losing the entire prject for sure? This is a classic case of missing the forest for the trees. The project needs to evolve or it is likely to die. Personally, I’m hoping they get the funding and can more forward, but if not it’ll be a shame if the whole project sinks over what amounts to petty stubborness.
So losing all the developers is a good idea?
> They probably use a cross compiler
Actually they don’t. They have a very strict no cross compiling policy except to bootstrap a new port.
tidux,
“Actually they don’t. They have a very strict no cross compiling policy except to bootstrap a new port.”
Apparently you are right, but I really can’t help but ask why? I guess it’s a strong sense of ideology. But in this case why not take the pragmatic route since the bytes produced are 100% identical regardless of where the compiler runs.
Theo says their machines run 24/7. If the project could afford it then that’s fine, but otherwise it seems logical to cut down the inefficiencies that burden the project with such a high reoccurring cost.
http://marc.info/?l=openbsd-misc&m=138990805222063&w=2
I think they’ve approached the problem from the wrong direction (how do we justify running these servers 24/7). They should be more receptive to suggestions for optimizing their energy consumption.
OpenBSD is not known to have best powermanagement code(if any). The money would be best spent by implementing all power-saving advances that other OSes have, instead of burning them through electricity bill. That would permanently cut their power consumption.
Stop supporting a dozen obsolete architectures to save some money
Edited 2014-01-17 08:39 UTC
The reasons it isn’t done are well explained in the e-mail[0] Aflman is citing above.[1] And may I add, Theo has a valid point there – these architectures ultimately help detecting bugs earlier then those would affect production systems, which is yet more important for a security-focused project.
[0] http://marc.info/?l=openbsd-tech&m=138973312304511&w=2
[1] http://www.osnews.com/permalink?580859
natural selection incoming
I totally agree. Drop the useless architectures.
Here is another idea to consider (very unpopular, I know): re-merge the OpenBSD and NetBSD projects.
What’s useless and who decides what’s useless? You?
No, the project does.
I think some of the privacy and digital rights groups should be made aware of the importanec of a security focussed project, and make the link between it’s ongoing success and their own goals.
Therefore I think they should contribute too.
I don’t know enough about the differences between them, but I hope the idea helps someone who does.
Wish I had some extra funds laying around I could throw their way. Sadly I don’t.
People who never installed nor know about the project are always full of ideas about how they should run the project.
“Drop architectures; make better energy saving code (on vax? alpha? really!?); use qemu; cross compile; merge with NetBSD; blah, blah, blah…”
They’re NOT asking for suggestions. They know, better than anyone else, what works and what not. They’ve run the project for almost 20 years, with very low resources and they have been very succesful.
If you can’t help, spread the world for who can and stop doing silly statments.
Unfortunately thats the nature of the beast. While I myself consider all input helpful, some others may not. When I needed funding I got plenty of advice from people alot of it didnt work, this current method I’ll let you know. But this isn’t about me its about OpenBSD. I think everyone right now is just trying to find a solution and I’m just surprised the two biggest corporations who use OpenBSD code, Apple and Microsoft are not assisting OpenBSD in any way. Microsoft uses it in UNIX subsystem or whatever its called now. But once again, I think it comes down to perception. People see what they consider a huge entity, and they see they have sponsors and they see they are in the news and they figure everything’s nice and dandy and sustainable and thats not always the case. You have to make people realize that you need funds to go on. Also, a good rule of thumb is not to threaten to take a free project and make it a paid product. You disorient your userbase, you make them angry and they will go elsewhere. Now, those that do stick around for your paid product, who think its reasonable then start to have certain expectations and you gotta offer them something to enhance the value, which in turn requires more resources. So, there are no easy answers and no quick fix. Im sure that the OpenBSD team does appreciate the donations that probably have come in, but now they have to find a way to sustain that. But, lets say that the BSD website has had 500 views and 500 people donated $5.00. Thats $2,500 dollars which goes to their goal. Now I know that doesnt seem like much but it gives them a good starting base. And yes for those of you that are probably wondering, I donated to them.
Roberto J. Dohnert
Head Developer
Black Lab Linux
http://www.blacklablinux.org
dbolgheroni,
It seems like you are referring to my posts, but I am not telling them how to run the project. It’s just an observation that there are optimizations to be had, and it seems self-evident that they cannot afford to be ignoring them.
They’ve made that *crystal clear* in their responses.
It’s worked in the past, but what happens if they won’t adapt for the future? Trust me when I say this, I’m being critical because I want them to succeed!
But what would the Internet be without backseat drivers and armchair generals?
Are we
reaching the oblivion
of secure communications
over x86 silicon?
:/
But when I do .. http://i.imgur.com/9fWXbpe.jpg
A popular suggestion has been to drop architectures that are not widely-used. The fact that a smaller BSD project like DragonFlyBSD only supports x86 and x86-64 might explain why it continues to hang on.
I do understand that they use multiple architectures for testing, but don’t understand why building binary packages cannot be done through cross compiling, using the fastest and most efficient one for this purpose (I assume that packages generated will be 100% same as those generated on native). They can run them later on respective architectures and do actual testing. Just why not to cross compile?
Why on Earth we try to reduce energy consumption at all, we can carry on without energy efficient technologies?! Lets follow OpenBSD developers logic, lets spend and ask someone else to pay for our “huge” bills.
I like OpenBSD, but their attitude in total refusal of reducing cost is childish and not healthy for project.
http://www.openbsd.org/faq/faq5.html#ProbXComp
From post containing “Many pre-built packages for each architecture: ” – i386: 7976, sh: 989.
It is possible that not all packages are possible to build or use under “sh” architecture, but that number would be bigger if cross compiled.
I do agree that building an OS in native way stresses a lot system and tests it, but will building actually be as good as running cross-compiled packages ~7976 and not 989. Greater number of software should be a better test then an OS building process due to variety of executed code and bugs! But as some of those architectures are ancient and except for few persons in world using them, having a lot of packages has no sense they won’t be able to use them all and the only easiest test is building OS natively. From http://www.openbsd.org/faq/faq5.html#ProbXComp : “Without native building, it is much more difficult to be sure that the various platforms are actually running reliably, and not just booting.” – For me it means : instead of building newer OS in 6 hours by cross compiling we will build it natively in 48(my guess) hours using a lot of electricity and will see if it works after another 48 hours of self building, instead of doing an average Joe system installation process 2 hours (maybe), configuring it (4 hours), installing additional cross compiled packages(4-8 hours) and actually using it for 5-6 hours, it will not be known whether an average Joe will be able to do previously described steps as the most important is self building, simply because there is no average Joe on that non-mainstream and ancient architecture to test it.
I have been reading that mailing list thread, the most logical thing is to drop old architectures, but due to developers will leave, have impression like someone stays in OpenBSD only to code for defined architecture, simpler solution is cross compiling regardless what they say, it can be done fully or partially. Or find funds or project dies, they refuse to adapt.
A BSD OS was my first experience w/ a Unix more than 12 years ago and it was a better experience than I made w/ Linux some time later…
When a project is economically not sustainable why should it not die peacefully? And with economically, I don’t mean necessarily money! I mean resources in general, the man power, the passion of people who work on the project and pay it out of THEIR own pocket without annoying others with funding requests.
There are many other OSs to fill the niche: Solaris, AIX and for those who want an OSS OS, can go with IllumOS or Linux. Let non-sustainable things die out, it’s no a big loss. Otherwise those things would not die out.
Where’s, for example, Sony as a major contributor? Since all recently released playstation 4 has an OS freebsd based?
Troll or ignorant?
OpenBSD != FreeBSD.
It appears that Sony has used code from both FreeBSD and NetBSD for the PS3 and PS4, but not from OpenBSD. Hence they can get around contributing any cash to OpenBSD’s ongoing financial crisis.
The biggest clues to the PS3 using a BSD-based OS come from the copyright statements that can be accessed from the PS3 menus. It’s pretty clear from the PS4 dev kits that “Orbis OS” draws heavily from FreeBSD 9.
Edited 2014-01-19 18:17 UTC
Bob Beck (beck@) has announced the Foundation has received over $100,000 in donations from users / companies in the past week.
http://marc.info/?l=openbsd-misc&m=139024400731106&w=2
A campaign page for 2014 has been setup to help the project throughout the year. It includes a fancy
progress bar.
http://www.openbsdfoundation.org/campaign2014.html