Hacker News new | past | comments | ask | show | jobs | submit login

That's the thing -- it's a testbed for a lot of experimental, some would even say unnecessary, technologies.

Just like Linux taught me to appreciate FreeBSD, FreeBSD taught me to appreciate OpenBSD. OpenBSD -- the world's simplest and most secure Unix-like OS. Creator of the world's most used SSH implementation OpenSSH, the world's most elegant firewall PF, the world's most elegant mail server OpenSMTPD, and the OpenSSL rewrite LibreSSL. OpenBSD -- the cleanest kernel, the cleanest userland and the cleanest configuration syntax.

Ref. http://marc.info/?l=openbsd-misc&m=139321387226212

http://itwire.com/business-it-news/open-source/62641-crypto-...




I would like to second this sentiment.

As someone who has been using FreeBSD professionally since 1998, and built the first VPS provider on it[1] and then built all of rsync.net on it, I am a huge proponent of FreeBSD.

However I (and others) have been frustrated for years by the culture of new and shiny that has displaced a culture of stability and long term support.

Rather than rehash it all here, let me suggest that if you are interested, you may read this mailing list posting:

http://lists.freebsd.org/pipermail/freebsd-hackers/2012-Janu...

... which I wrote in January 2012. Much of it remains true today. That being said, when we reworked the rsync.net platform from the ground up in late 2012, we did indeed go forward with FreeBSD.

[1] JohnCompanies "Server Instances", based on FreeBSD jail, were made available in Fall of 2001. A year or so later the world settled on "Virtual Private Server" as the de facto descriptor.


it's a testbed for a lot of experimental, some would even say unnecessary, technologies.

Nope, that's DragonFly BSD. FreeBSD is far more conservative in comparison, occasionally integrating some more daring idea. In contrast, NetBSD used to silently innovate all the time (and I think still do, what with rump kernels) and even OpenBSD, despite being rather vanilla on the outside, is internally innovative in the area of security.


I think it varies. FreeBSD does have a lot of stuff, like netmap, capsicum, that is quite experimental in some sense, although capsicum is a great design. NetBSD has rump kernel, but thats not much code, it is just(!) portability done right, some build fixes and a little code.


You said a lot of nice and true things about OpenBSD, but you didn't back up your claim against FBSD: What unnecessary and experimental technologies are you referring to?


He's just an OpenBSD fan. He's entitled to his opinion, even if he's wrong.


I didn't read the part where his entitlement to his opinion was questioned. The question I read seemed pretty straight forward and reasonable.


Although it turns out he's not.


Can I write ls "é*" in an openBSD console at my first boot? Else, why is it unclean to have accent by default in a shell?


Testbed may be a strong world, but there's definitively a lot of features thrown into FreeBSD and maybe more than there should be. Configuration syntax for all the different pieces are all over the place, it clear that there's not really anyone in charge of FreeBSD. That might be a good thing, and the reason why FreeBSD has all these feature in the first place. It's just really confusing to new comers, it unclear which of these feature I need, and which are optional. Even configuring drives are a bit of a mess, do I need GEOM, softraid, is GEOM a prerequisite for softraid, should I go for ZFS, how much space do I need? I've installed FreeBSD only to find the keyboard to be weirdly mapped, OpenBSD was as you would expect.

For performance, FreeBSD beats OpenBSD in all cases. The thing is, many of us would do just fine on OpenBSD while finding configuration and installation easier and the whole experience of using the system more consistent.


For performance, FreeBSD beats OpenBSD in all cases. The thing is, many of us would do just fine on OpenBSD while finding configuration and installation easier and the whole experience of using the system more consistent.

I really liked and used OpenBSD in the 2000-2002 timeframe (then I returned to Slackware, later OS X). At some point, hardware support became an issue. How good is it these days?


About the same.

It wouldn't detect my SATA HDDs on a GA-945GCM-S2C board (no issues with that on Windows/OSX-Hackintosh/Debian/FreeBSD), had to use USB-HDD to install. Only option for my nVidia graphics was the nv driver, which is actually many times slower than the vesa driver somehow. And vesa lacked any widescreen modes, so video was practically unusable for me. Other than that, everything worked out of the box. LAN (Realtek), audio (HDA), etc.

Only thing I didn't like on the software side (but many people would prefer) was that it's really bleeding edge. I much prefer Xfce, but still rely on some Gnome apps. OpenBSD is much further along with Gnome 3, so a lot of apps like gedit, gnome-system-monitor, gcalctool, evince etc start to stand out like a sore thumb. That'll come to FreeBSD too, but at least I have a bit more time to find alternatives.


Hardware support is rather good. Nvidia graphics chips are an issue, other than that I wouldn't expect to many issues.

Laptops are harder than desktop, mostly due to ACPI and I still trying to get a multitouch trackpad to work (I'm missing a right button).

Linux and FreeBSD has better hardware support, but when stuff is supported in OpenBSD it just works. Only exception I ran into is wireless card in an old MacBook Pro which should work but doesn't.


> Only exception I ran into is wireless card in an old MacBook Pro which should work but doesn't.

If this was a bwi(4) card, the problem is probably fixed in 5.6: http://marc.info/?l=openbsd-cvs&m=140585757928895&w=2


Thanks, but it's an athn: http://www.openbsd.org/cgi-bin/man.cgi/OpenBSD-current/man4/...

I've seen others get the same error, funny enough the man page says: athn0: device timeout A frame dispatched to the hardware for transmission did not complete in time. The driver will reset the hardware. This should not happen.


Yes, that's known. Coincidentally, this came up today: http://marc.info/?l=openbsd-misc&m=141501277608157&w=2

The device timeout message might not necessarily be a fatal problem. If it doesn't have an impact on actual usage you can just ignore it. Otherwise, please file a quality bug report so we can learn more about it.


To chime in bit on FreeBSD, going Linux -> FreeBSD can be relatively painless, and FreeBSD will be a good place to stop if there is no further need of performance/purity/cleanness.

The was a time it was the only blessed BSD distribution for Java, unicode support has been pretty good for years, the ports management is easy and forgiving, documentation is generally clean and complete, help resources have a polite and professional feel.

In general there are stronger alternatives for a production server (OpenSBD definitively) but it's a very nice environment for a workstation or development server.


FreeBSD is an excellent production server, and I use it all the time. And for "non-hardened" and "protected" production servers, the performance numbers went solidly to FreeBSD for my use cases. Modifying the kernel from GENERIC to something specific to one's needs also helps.

That said, I'm also OpenBSD fan, using it as a "hardened" (Internet facing) production server, with minimal ports, for things like Postfix, DNS, etc. OpenBSD is also well used in network management/control as a router, firewall, etc.

Each has its use case, and does quite well in them.


Experimental ideas are almost always shot down by the developers




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: