FreeBSD 8.4 Released 80
kthreadd writes "The FreeBSD project has released version 8.4 of the free operating system with the same name. Highlights of this version include GNOME 2.32.1, KDE 4.10.1. In this release, focus has been put on improving stability and storage capability. The ZFS filesystem has been updated to support feature flags for ZFS pools, asynchronous destruction of ZFS datasets, LZ4 compression and ZIO NOP-write optimization. Also, support has been added for all shipping LSI storage controllers."
last (Score:0)
last post
Re:last (Score:2)
No, me!
More seriously, as a ZFS user (first on Solaris, now on Linux), I'm happy that ZFS outgrew Oracle and lives on as a community opensource project. I greatly prefer it to btrfs, if only on a management level. And ironically, ZFS is now completely free of Oracle influence, while btrfs isn't.
Netcraft? (Score:0)
Amazing that this was posted several hours ago and there's no "Netcraft Confirms" postings...
Seriously. FreeBSD rocks.
only recommended if you need to stay on 8.x (Score:5, Informative)
Most desktop users won't want to install this release. FreeBSD 9.1 was released in December 2012, and is the most recent stable release. This 8.4 release is a point release in the still-maintained 8.x series, intended for people currently running 8.3 who for one reason or another don't wish to upgrade to 9.x yet, but who do want an incremental upgrade.
Re:only recommended if you need to stay on 8.x (Score:4, Interesting)
Except 8.4 has:
Better hyperthreading support than 9.1
Newer ZFS features than 9.1
Better snd_uaudio and snd_hda audio drivers than 9.1
These things were MFC'd to 8-STABLE and 9-STABLE after 9.1-RELEASE, so 8.4 is really a better release an some aspects than 9.1 is.
Re:only recommended if you need to stay on 8.x (Score:3, Informative)
Re:only recommended if you need to stay on 8.x (Score:2)
so basically you are saying I need to do a build world after I drop 9.1 on the new pc tomorrow, thanks.
Re:only recommended if you need to stay on 8.x (Score:2)
Except 8.4 has:
Better hyperthreading support than 9.1 Newer ZFS features than 9.1 Better snd_uaudio and snd_hda audio drivers than 9.1
These things were MFC'd to 8-STABLE and 9-STABLE after 9.1-RELEASE, so 8.4 is really a better release an some aspects than 9.1 is.
And an update to the 9.x branch bring up all those features will make 8.4 a wash for 9.x users. It's great news for 8.x branch users.
Re:only recommended if you need to stay on 8.x (Score:0)
How much better are these drivers? I tried FreeBSD 9.1 some time ago, and the sound was so bad I just went back running to Linux.
Re:only recommended if you need to stay on 8.x (Score:0)
It's not about how it sounds, these drivers were unfinished or terribly buggy. The OSS USB audio driver for example is a joke, you weren't supposed to plug in or remove any USB audio devices while the driver is running because it would crash the OS. Yes, seriously.
The HDA driver doesn't actually read HDA configuration data and set up your device. If it's not one of a handful of well known device you're responsible for manually creating a text file that configures the device properly.
So maybe FreeBSD 8.4 fixes some of this, bringing it into line with operating systems from 3-5 years ago.
Re:only recommended if you need to stay on 8.x (Score:0)
While I'm using 9.1 on my personal machines, I appreciate that they still provide support the 8.x release branch. Supporting the legacy releases definitely make FreeBSD an attractive option for enterprise environments.
Re:only recommended if you need to stay on 8.x (Score:2)
I love the legacy support even for my home machine. I set up my server 2 years ago with the 8.x series, and their continued support makes it easier to maintain than if I had to upgrade to 9.x. For a new build, I'd definitely go with the 9.x series.
Re:only recommended if you need to stay on 8.x (Score:0)
It's great that they support older versions.
Allowing a version to grow to full maturity is very important.
It makes for a really solid lineage.
I have servers still running older versions going back to 6.x, that due to late updates,
never needed to be upgraded, and are able to continue on in their 6.x awesomeness.
Obviously there's advantages to switching over to the latest and greatest,
just in those cases there was no need, due to the late upgrades.
When you spend a lot of effort modifying a system to your needs,
it can be upsetting and painful to have to have to do it again,
for an updated system.
Love you BSD,
HasH @ TrYPNET.net
Re:only recommended if you need to stay on 8.x (Score:2)
Most desktop users won't want to install this release.
How many people run FreeBSD on their desktop? (No, I'm not counting OSX.) The big selling point of FreeBSD is its robust support for ZFS, which makes it great for a file storage server. But it's an extremely marginal and weird choice for a desktop environment.
Re:only recommended if you need to stay on 8.x (Score:0)
I'm a weird margin!
Re:only recommended if you need to stay on 8.x (Score:1)
Re:only recommended if you need to stay on 8.x (Score:1)
Re:only recommended if you need to stay on 8.x (Score:2)
Both Linux and FreeBSD allow building exactly what you want from source if you want to. It's not quite right to say one gives more control over what goes into the system. FreeBSD has a different set of trade-offs in how things are packaged, and their default choices and packaging distribution choices are nice for some purposes. But ultimately there's nothing you can match in multiple Linux distributions if you feel like it. There's always Linux from Scratch if you're hardcore about controlling what goes into your system.
Re:only recommended if you need to stay on 8.x (Score:3)
The difference between Linux and *BSD has never been wider. I love BSD because of how /etc is configured. Linux made a mess of things and with massive scripts and subdirectories. Things have changed so much since the 2.0 kernel days its almost an entirely new OS.
Re:only recommended if you need to stay on 8.x (Score:2)
Re:only recommended if you need to stay on 8.x (Score:0)
I've been using FreeBSD on my desktops, laptops, and servers since late 2001. I've never had issues with it other than ACPI support on the laptops being hit-or-miss. And the ports collection makes package management trivially easy.
Re:only recommended if you need to stay on 8.x (Score:2)
"How many people run FreeBSD on their desktop?"
No one really knows for sure. I count as 1.
If (and only if) your hardware can run it, you're better off running a FreeBSD desktop than Linux. If your hardware is incompatible with FreeBSD, then run Linux. If the hardware won't run Linux, then either run Windows or buy new hardware. (Don't run Windows). As long as you're buying new hardware make sure you can run FreeBSD on it.
Re:only recommended if you need to stay on 8.x (Score:2)
Call me crazy, dumb, or just stuck with the Linux way, but I think I'd have to re-learn a decent amount to be able to successfully run and maintain FreeBSD on my desktop. Sure, I learned many of the UNIX basics of the command line before switching from Windows, but I think there are a lot of things that would seriously stump me. For example, things like "mount -o loop" to mount an image file as a loop device and "free -m" to get a quick reading of memory usage are much different, and not exactly easier. I just couldn't get used to the package management tools, and I'm no guru at compiling (in fact, I hate it...).
It would be nice if Linux distributions were like this though; 8.x has a legacy base OS with what seems to be extremely up-to-date packages otherwise. With Linux it's basically either all (Arch, Gentoo) or nothing (Debian, CentOS) as far as updates to the non-system packages go. It seems that openSUSE does major upgrades on just a few regular packages (mainly web browsers) and I remember reading that Ubuntu was going to too, but it's nothing compared to what FreeBSD does. FreeBSD's support really is good... and I heard that its separation of system and userland makes complete system upgrades work well. That said... I intend to someday learn it better to be able to use it. One of these days I intend to figure it out...
I tried PC-BSD, but I'm not too much of a fan of the way it does some of the things it does... not to mention it comes across as incredibly bloated to me. In theory, I like the way plain FreeBSD is set up.
Re:only recommended if you need to stay on 8.x (Score:2)
Re:only recommended if you need to stay on 8.x (Score:2)
I tried and failed yet again to get a full KDE desktop up and running. The directions I used--a page on the FreeBSD site found through a web search--simply said "pkg_add -r kde4". Nope, nothing. Did it actually install X.org? Didn't seem so, so I logged back in as root and did that. Logged back out and in again as my user, and entered startx: returned endless "failed to load module" warnings, no drivers available, no screens found, unable to connect/connection refused, blah blah blah. Before following the directions from the same page to get X to load right into KDE... startx would in fact start X with twm, but... there was absolutely no keyboard *or* mouse support once I was in X.
Seriously, what the hell am I doing wrong?! I can use Slackware, Debian, CentOS--hell, nearly every major Linux distribution aside from Arch, Gentoo and Crux... but when it comes to the BSDs I can never get anywhere.
Re:only recommended if you need to stay on 8.x (Score:2)
ah me?
Re:only recommended if you need to stay on 8.x (Score:0)
aparantly in BSD land, they haven't had to deal with the GNOME 3 headache that has been plauging the linux comunity for the last 4 years.
mmm gnome 2.32, thats a decent reasons to jump ship
Re:only recommended if you need to stay on 8.x (Score:2)
Why are they doing 8.x? (Score:2)
Re:Why are they doing 8.x? (Score:2)
They typically keep putting out point releases in a series for about five years after the initial .0 release, so at any given time the current and previous one or two series are supported. But they eventually get phased out, e.g. the last 7.x release was 7.4, which came out in early 2011 and stopped being security-managed in early 2013. Wikipedia has a timeline showing the release/support history [wikipedia.org].
One of the reasons for maintaining the legacy branches for a few years is that, within each series, FreeBSD commits to maintaining binary compatibility. So, upgrades are simple & quick and won't break any third-party software you've built yourself from source. There's a bit more description of the difference in section 25.2.3 here [freebsd.org].
Re:Why are they doing 8.x? (Score:1)
FreeBSD will have official branches as long as it has volunteers. There are companies who have products heavily invested in 6.x and 7.x and the answer has always been something like "you're more than welcome to resurrect those branches, become the official maintainer, and continue to backport improvements but we don't have time or resources"
So if someone in the community steps up, we could see newer 6.x and 7.x releases as long as people are OK with the ports tree ignoring their existence. The modern ports tree simply cannot support every OS version because it's quickly evolving and improving at quite a rate and some of the new requirements simply don't come with the base systems of FreeBSD versions before 8.x.
Re:only recommended if you need to stay on 8.x (Score:0)
Most desktop users won't want to install this release.
Most desktop users probably want to check out PC-BSD, which builds on top of FreeBSD.
Re:only recommended if you need to stay on 8.x (Score:0)
If by "builds on top" you mean it makes FreeBSD so bloated it makes Windows seem frugal comparison, then I agree wholeheartedly.
FP (Score:-1)
First Post
Legacy Release (Score:4, Informative)
Re:Legacy Release (Score:0)
Still it does have some fixes that are not available in 9.1-Release.
Yabut... (Score:2)
It took thirty years, but they finally came out with a system twice as good as the old original BSD 4.2 (no open/free/net) that I used back in '83!
Desktop environments (Score:1)
If memory serves OpenBSD 5.3 was released with GNOME 3 and KDE 3.5. Here we see FreeBSD 8.4 released with GNOME 2 and KDE 4. Can anyone shed some light on why one BSD operating system has a modern KDE and outdated GNOME desktop while the other has the reverse?
Re:Desktop environments (Score:0)
Because Gnome 2 is usable while Gnome 3 is not (yet, I hope).
Re:Desktop environments (Score:3, Informative)
If memory serves OpenBSD 5.3 was released with GNOME 3 and KDE 3.5. Here we see FreeBSD 8.4 released with GNOME 2 and KDE 4. Can anyone shed some light on why one BSD operating system has a modern KDE and outdated GNOME desktop while the other has the reverse?
Because FreeBSD 8.4 is a historical maintenance release for the FreeBSD 8 series; FreeBSD Release is at 9.1-p3.
Re:Desktop environments (Score:0)
No Gnome is part of the ports collection, which is the same for both 8 and 9. So you always get the latest version that is in ports no matter which version of FreeBSD you're on. It's only the core of the operating system that is actually part of FreeBSD the distribution.
Re:Desktop environments (Score:2)
Re:Desktop environments (Score:1)
openbsd is close to having kde 4.8
Re:Desktop environments (Score:0)
Actually 4.10.4 is what's in ports and being worked on right now.
Re:Desktop environments (Score:2)
that's amusing, to judge an OS mainly used for servers and appliances by what desktop version it has. If it's that important by all means choose your BSD by the desktop you like as long as the devices you want are supported. I myself prefer xfce4 which runs on them all
Please if some FreeBSD dev sees this... (Score:0)
could you please guys implement some anti exploitation technologies such as ASLR out of the box ? Or maybe dedicate a manpage explaining the dev team views over such matters.
I've been a long time user of FreeBSD and i can't help but to feel it keeps dragging behind in this field.
Or please someone explain me why i shouldn't be worrying about that.
Re:Please if some FreeBSD dev sees this... (Score:1)
could you please guys implement some anti exploitation technologies such as ASLR out of the box ? Or maybe dedicate a manpage explaining the dev team views over such matters.
I've been a long time user of FreeBSD and i can't help but to feel it keeps dragging behind in this field.
Or please someone explain me why i shouldn't be worrying about that.
Explanation: You're running BSD. What gains is ASLR supposed to provide you, the end user? Protecting you against a custom crafted process injection attack on your specific BSD distro written by someone who could have made more profit creating a bogus kickstarter campaign or by attempting to mine bitcoins?
(I kid... there's enough stuff running FreeBSD that no ASLR by default is a bit odd... but it sure makes debugging easier)
Re:Please if some FreeBSD dev sees this... (Score:0)
Thank you for your answer. I can't decide on which side of irony you stand. I'd prefer if your stance was the one between parenthesis.
As for the first part that was a legitimate question, i don't use FreeBSD as a desktop, i run a server which got through all the upgrades since FreeBSD 5.2, it has a few jails and paranoid setup. And most of all i use FreeBSD because i expect it to be relatively on par with OpenBSD for security.
And as an ROP mindfuck lover myself i was really disappointed when the Intel SysRET vulnerability came out and the exploit for freeBSD was a straightforward piece of cake compared to the Windows7 flavor. Thus some pretty random guy could just have owned my whole installation more easily then a desktop Microsoft product, bypassing the jails and everything at the same time, i almost shat myself.
Well if anyone involved reads this, please try to make the situation better if it deserves to be :p
Re:Please if some FreeBSD dev sees this... (Score:0)
There are diminishing returns for things like chroot, jails, and all the crap that Linux has. Most kernel exploits will bypass all of it, and of course the more complex "security" features you have, the more kernel code required and the more opportunity for exploitable bugs to creep in.
This is why OpenBSD has rejected all of these things. Existing capabilities, such as chroot, are good enough. FreeBSD is not much secure than Linux. Both platforms have many features, and continually keep piling more on top. While Linux kernel exploits are a semi-monthly occurrence, FreeBSD has at least one or two every year. Compare that to OpenBSD, which averages significantly less than 1 per year.
OpenBSD isn't more secure because it has better developers, it's more secure because it has fewer developers committing fewer features, and they're very conservative about new code. Even though everybody thinks OpenSMTP is hot shit, it'll still be several releases before it becomes the default mailer.
Re:Please if some FreeBSD dev sees this... (Score:2)
ASLR [wikipedia.org] has been usefully complicating enough vulnerabilities to have proven it's worthwhile. At this point it's quite near being an industry standard for any system that follows good security practices. It's really not credible to reject it anymore as too complicated to risk bothering with. Yes, some of the issues can be addressed more deeply, too, but security should be layered and redundant.
A major cause for why there are less exploits on the *BSD kernels is that the rate of innovation is so low. New and rapidly changing code is where a lot of security bugs come from. If Linux kernel development did a lot less work each year and has a smaller feature base, they could spend a lot more time on hardening too. The *BSD approach isn't a bad one, but it's a trade-off with good and bad sides to it.
Re:Please if some FreeBSD dev sees this... (Score:0)
Re:Please if some FreeBSD dev sees this... (Score:3)
If you read about ASLR and I recommend the OpenBSD paper on the topic (I think they implemented it 1st) the technique only makes it more difficult, it doesn't SOLVE the problem. If you benefit from ASLR, then you have a security problem already. I'm not sure I like the idea of losing entropy and slowing down malloc for something that does little to stop attacks from rogue processes (which can be compromised by other methods.)
Sure, I might prefer you DoS some process by crashing it repeatedly instead of taking control over it... but I would prefer you not corrupt or crash it at all and I would want more effort put into better jails and damage control. Removing strcpy and other common trouble makers like OpenBSD does would be nice. ASLR adds to the level of complacency (except on OpenBSD where paranoia is expected.)
The goals should be what they always were. New buzzwords be dammed.
NOTE: I'm years behind on my BSD.
I'm not convinced we shouldn't be moving towards microkernel hybrids like Darwin and towards a full microkernel. We took a big speed loss going to protected memory management which was accelerated in hardware and new CPUs made the transition almost unnoticeable. Maybe we should be aiming for something similar? (just isolating most drivers would help; you could leave the FS and HD within the kernel.) I also hate to think of how cool it could have been in Multics was used in place of unix... every unix has been more bloated for decades and I would love to swap RAM, CPUs, etc without rebooting. I find netBSD's work on including an interpreter in the kernel to be interesting in a shocking kind of way. Anyhow, the point is that ASLR is just a niche band-aid and not the most important feature we "must have" to continue living... and we've survived for decades without it. Such complaints sound more like an IT person speaking.
Re:Please if some FreeBSD dev sees this... (Score:0)
tcsh (Score:0)
I'm boycotting FreeBSD until such time as they change the default root shell to ksh or something Bourne compatible. There's just nothing more frustrating then tcsh.
Re:tcsh (Score:3)
Seems a stupid reason, because it takes all of 5 minutes to "fix."
Re:tcsh (Score:0)
I have not checked lately, but it used to be the case that a lot of things broke if you changed the root shell. Maybe they have fixed that.
As long as it's not bash I'm happy. Bash is terrible, why would you ever use that?
Re:tcsh (Score:2)
Re:tcsh (Score:1)
The ports tree has Bash 4.2.45, i'm not sure how up-to-date this is compared to linux.
As for a list of shells, the Ports tree reports 49 different shells, although some of them are just tools: http://www.freebsd.org/ports/shells.html [freebsd.org]
As for your question, FreeBSD does has KSH and ZSH.
Re:tcsh (Score:1)
As an addition to my previous comment. Upon further investigation (by which I mean I discovered bash --version), I found that the version of Bash in the ports tree is indeed a GPL V3 version of Bash. I assume this means that whilst the FreeBSD project can not use any GPL V3 code in the operating system itself (I believe by FreeBSD 10 they want to have removed all GPL code full stop), there is no restriction on the licencing used by software in the ports tree (within reason).
Also, I quickly checked my Debian Wheezy box and that runs Bash 4.2.37, and Arch is using the same version as FreeBSD. I guess it shows that depending on the package, the ports tree can be rather bleeding edge at times.
Re:tcsh (Score:2)
That is correct. They don't want GPLv3 in the operating system, but ports is OK. FreeBSD is very different from many Linux distributions in the sense that the operating system and most third party software is managed separate from each other. In Linux you often have a single package manager that handles both the operating system and all your software, with no clear distinction between them. In FreeBSD these are completely separate.
Re:tcsh (Score:2)
I have not checked lately, but it used to be the case that a lot of things broke if you changed the root shell. Maybe they have fixed that.
I think the "fix" for that is toor. You can change its shell to whatever you like and use it instead of root. Although I've never tried it; I just have an alias that calls sudo -i exec zsh -l, which seems to work great.
Re:tcsh (Score:2)
I don't believe you've checked... EVER. There's nothing TO break by changing the root user's login shell.
All the services that start up (eg. crond) are completely ignorant of what root's shell is set to. Services don't go through the process of logging-in as root and then spawning processes... They use the setting and environment they inherited from the rc* scripts, from start-up, while the system was single-user.
Any programs or shell scripts you run once the system is multiuser as root, will be run with whatever interpreter they're set to use with the #! line. They don't know or care what your shell is.
In short, there's practically no way that changing root's login shell to something different could possibly break anything. Now, if you do something stupid, like replace /bin/sh with something else, you'll break the whole damn system, but that's an idiotic mistake, and not how anybody should ever change a user's preferred shell.
Re:tcsh (Score:5, Funny)
There's just nothing more frustrating then tcsh.
You have clearly never been forced to use Lotus Notes.
Re:tcsh (Score:0)
A meme is born.
Re:tcsh (Score:3)
I'm fucked. I use tcsh on my *nix boxes and now my employer is moving to Lotus Notes and I have to move all the Exchange shit over to Domino and join it to the French company that bought us.
Re:tcsh (Score:2)
I'm fucked. I use tcsh on my *nix boxes and now my employer is moving to Lotus Notes and I have to move all the Exchange shit over to Domino and join it to the French company that bought us.
I am told by those who know more than me that Domino is actually pretty good, it's the client that sucks dead dingo kidneys. However, the Exchange to Domino migration has been known to trigger a wide variety of psychological and medical disorders.
Re:tcsh (Score:0)
I'm boycotting Linuxs until such time as they change the default shells to tcsh or something csh compatible. There's just nothing more frustrating than bourne.
-HasH @ TrYPNET.net
Re:tcsh (Score:0)
omg dats racis
Re:tcsh (Score:2)
Huh. I change all my shells to tcsh. I set up my configuration the way I like it in 1996 or so, so it's the path of least resistance for me.
Re:tcsh (Score:0)
I love tcsh and find it mildly amusing when linux users get really confused when I tell them I'm not using bash.
Should a *BSD newbie install this at home? (Score:3)
Myself am a windows user with a basic++ linux knowledge: I know how to install and update a distro(I prefer debian based but right now am wanting to test Fedora 18), compile some packages from source, has poked some kernel compilin', made some kde translation on the past, reported some bugs on FOSS software, etc. Now I have my Dell laptop with Windows 7 and I'm planning to back it up and format it and I'm thinking to put it up Fedora 18, but the BSD world has intrigue me and I've made some test on virtual machines.
BTW I'm primarly a Java developer how some times made some tiny personal project on my laptop and who enjoys a good Quake 3 match(for remembering my old days on the College fragging like there were no tomorrow) but who actually prefer to enjoy the time with my wife and the kids.
Could a user like me install some BSD distro and used it regularly to this basic things I've listed? if so, which BSD do you suggest?, bonus points if lastest KDE's it's available
Thanks for your suggestions!
Re:Should a *BSD newbie install this at home? (Score:0)
I'm sure you've gotten the obligatory PCBSD suggestion here and there enough times that you are aware of what it is. That distro has worked well for me to get a quick desktop up and going and would work for you as well. I think you are all good to go except for the Java development part- which will work as long as you sign your soul away in triplication and know what your dev environment is going to be before jumping into it. It can get messy pretty quick if stuff gets left around everywhere but definitely is quite possible to do Java development on a BSD box. Linux apps which rely on java can work in linux compatibility mode too.
Re:Should a *BSD newbie install this at home? (Score:2)
Re:Should a *BSD newbie install this at home? (Score:2)
Re:Should a *BSD newbie install this at home? (Score:1)
No, your wife deserves better, check out Linux Mint instead, the 13 edition with mate esp. Everything works out of the box, and it's supported to 2017.
Re:Should a *BSD newbie install this at home? (Score:2)
I'd go for 9.1 though - the 8.* series is basicly bug fixes for older applications.
drivers (Score:0)
Last summer I was building a router with pfsense, only to discover that wireless drivers for 802.11 N were not included in the base BSD system? That was enough to change course entirely.
Please be sure to read the Release Errata! (Score:0)
The 8.4-RELEASE Errata notes are here:
http://www.freebsd.org/releases/8.4R/errata.html [freebsd.org]
It's of the utmost importance that all those considering upgrading (or installing) to 8.4-RELEASE read this document, particularly sections (3) and (4).
The issue with Intel NICs driven by fxp(4) is/was a very hot (heated) topic on the mailing lists [freebsd.org] (note: long thread, but very informative), and the issue described there may impact other NIC drivers as well. There is no workaround at this time other than avoiding DHCP (assigning static IPs in /etc/rc.conf).
Scheduler? (Score:0)
Is it worthwhile yet to use the 'new' scheduler?
Still harboring those binary blobs in the kernel? (Score:0)
I don't mean drivers but firmware... Not so free after all I guess.