Slashdot Mirror

← Back to Stories (view on slashdot.org)

FreeBSD 11.0 Released (freebsdfoundation.org)

Posted by EditorDavid on from the going-up-to-11 dept.

Long-time Slashdot reader basscomm writes, "After a couple of delays, FreeBSD 11 has been released. Check out the release notes here." The FreeBSD Foundation writes: The latest release continues to pioneer the field of copyfree-licensed, open source operating systems by including new architecture support, performance improvements, toolchain enhancements and support for contemporary wireless chipsets. The new features and improvements bring about an even more robust operating system that both companies and end users alike benefit greatly from using.
FreeBSD 11 supports both the ARMv8 and RISC-V architectures, and also supports the 802.11n wireless networking standard. In addition, OpenSSH has been updated to 7.2p2, and OpenSSH DSA key generation has been disabled by default, so "It is important to update OpenSSH keys prior to upgrading."

121 comments

  1. FreeBSD 11? by Anonymous Coward · · Score: -1

    It's FREE?

    Cool story brah.

    1. Re: FreeBSD 11? by Anonymous Coward · · Score: -1

      It's free until you factor in all the overhead of using a dead OS.

  2. FreeBSD is for losers by Anonymous Coward · · Score: -1

    and real men and women us LINUX. Linux is the only operating system that matters, and all other OS'es are for lusers! I said it! If you aren't using linux you are a LUSER!!!!!

    1. Re:FreeBSD is for losers by armanox · · Score: 0, Flamebait

      Tell me another good joke. Modern Linux is a PITA for system administrators, thanks to groups like Red Hat pushing cancers like systemd into the Linux ecosystem.

      --
      I'm starting to think GNU is the problem with "GNU/Linux" these days.
    2. Re:FreeBSD is for losers by rubycodez · · Score: 4, Interesting

      You know GNU/Linux is jumping the shark when THE major distro didn't take the time to research how to properly go from one LTS to the current one that has systemd and it clobbers the systemd related things. And Debian, that used to the the best engineered distro, just puts in needless circle-jerking complexity, even the /etc/motd is dynamically generated and the twats made the wrong choice of having a static /etc/motd.tail instead of the obvious solution of motd being static as per normal Unix convention and a motd.head to be generated. GNU/LInux is circling the drain

    3. Re:FreeBSD is for losers by Anonymous Coward · · Score: 0

      You know GNU/Linux is jumping the shark when THE major distro didn't take the time to research how to properly go from one LTS to the current one that has systemd and it clobbers the systemd related things.

      That's odd. I haven't had any issues with Linux Mint.

    4. Re:FreeBSD is for losers by rubycodez · · Score: 1

      Mint is my primary desktop: mint makes their own (modified from ubuntu) updating system. I wasn't talking about mint.

      the issue is with Ubuntu, 14.04 LTS to 16.04 LTS

      For my main home machine staying at 17.3 for now because no systemd suck. I have 18 at work and the systemd constipates the machine, what garbage

    5. Re:FreeBSD is for losers by Anonymous Coward · · Score: 0

      I assumed you were talking about Linux Mint (based on Ubuntu OR Debian) because of the "THE major distro" bit, seeing as Mint is the most popular Linux-based PC OS out there, with Debian a distant second and Ubuntu (modified Debian) in third place.

  3. I view of the last several weeks' history by fnj · · Score: 1

    Best wait a few weeks to make sure it's really released for real this time. This has been an embarrassing episode.

    1. Re:I view of the last several weeks' history by Anonymous Coward · · Score: 1

      The FreeBSD project never released version 11.0 before. Some early images were uploaded, tested and replaced. Anyone who is at all familiar with operating systems knows not to install and used unannounced ISO files. What is embarrassing is apparently some people are too impatient to wait for a release to be tested and announced before trying to use it. I have zero sympathy for them.

    2. Re:I view of the last several weeks' history by Anonymous Coward · · Score: 1

      /second. FreeBSD is deliberate. A few people jumped the gun. Hell even I did, when I saw the new images show up last night. I took a chance and figured they were official. Quite frankly, there are operations that would have just pushed the new release out, this late in the game. You have to commend these guys for doing it right. The guys who want to jump the gun and get burned are just whining that they guessed wrong.

    3. Re:I view of the last several weeks' history by 0100010001010011 · · Score: 1

      Its the open source equivalent of downloading and burning a "Golden Master.

    4. Re:I view of the last several weeks' history by Anonymous Coward · · Score: 0

      Someone mod parent up.

      How is this embarrassing? Because FreeBSD saw issues and decided "Hey, let's not release buggy code, let's rather delay the release" ???

      Good luck with your Ubunturd release on date given months in advance which in 100% of cases results in buggy code so "everyone know not to install but wait for X.Y.1 to come out" a few months later. Yeah, much better!

  4. Focus by Anonymous Coward · · Score: 2, Insightful

    The new features and improvements bring about an even more robust operating system that both companies and end users alike benefit greatly from using.

    I like these guys. They know what's important to focus their attentions.

    Others, well let's say they are more concerned with bells and whistles and eye candy.

    1. Re:Focus by arth1 · · Score: 2

      With that-which-should-not-be-mentioned killing linux, jumping ship to BSD becomes more and more attractive.

      Unfortunately, there are still some pieces missing, like lack of high performance file systems (zfs is high functionality, not high performance) and lack of software for many RAID controllers, or kernel-hooked utilities we take for granted these days like inotify, selinux and cgroups (when used by the admin to control resources, not as a crutch for that-which-should-not-be-mentioned).

    2. Re:Focus by Anonymous Coward · · Score: 0, Informative

      I dunno man, is there something better than ext4 out there for linux? 'Cause I don't see ext4 as being a better filesystem than ZFS by any metric.

      RAID controllers, I don't know the coverage.

      With respect to security and partitioning, FreeBSD has Capsicum, and has had Jails for decades. Oh, we also have Xen and Docker now, if you absolutely must. :)

      You might want to consider revisiting FreeBSD again; it's generally a well-thought-out design and implementation. There's not a 1-to-1 feature partity with Linux, but sometimes, that's the point.

       

    3. Re:Focus by dbIII · · Score: 1

      and lack of software for many RAID controllers

      Not such a huge deal now since ZFS exists to do the RAID calculations.
      RAID controllers just do not have much processing power or memory so getting the systems CPU to do the work almost always gives you better performance once you are talking about RAID6 or raidz2. I'm not even sure that they will give you better performance with mirroring.

    4. Re: Focus by Billly+Gates · · Score: 1

      I told my IT department something similar and they were on the floor laughing at me for my supposed stupidity.

      They use raid cards even for their home media servers as appearrently it is still believed the CPU does IRQ for each read or write. I explained pita and Dma solved that in 1997 and a modern CPU has I/0 in the CPU.

      My co-workers raid card at home was a windows 7 system too. Oh and he had ssds so no trim on that either! .... Sorry for rant but software raid has a very very bad wrap. People still feel it's 1998 still unless we are wrong

      --
      http://saveie6.com/
    5. Re: Focus by dbIII · · Score: 1

      I told my IT department something similar and they were on the floor laughing at me for my supposed stupidity

      They are very much out of date. Ask them what processors are in the RAID cards, they will not have a clue but they may look things up and get up to date.

      A current Xeon can do a bit more than a PowerPC CPU that came out eight years ago. ZFS (and several other things) do exactly the same RAID calculations as on those cards only on significantly faster hardware.

    6. Re: Focus by TheRaven64 · · Score: 1

      Add to that, recent x86 chips can do XOR in the DMA engine, so you actually have hardware RAID built into your CPU with a direct connection to the DRAM. Replacing it with one that's dangled off the PCIe bus doesn't sound too sensible.

      --
      I am TheRaven on Soylent News
    7. Re: Focus by swb · · Score: 2

      I generally agree that software RAID is fine, but will counter that even on generic Dell boxes with PERC RAID-6 controllers and SATA disks in RAID6, it's pretty easy to see writes do 250 MB/sec.

      The reality is your disks are off the PCIe bus no matter what.

      I don't think that it's controller performance per se that's the limitation anymore, it's that software defined RAID usually has a much richer feature set, whether its SSD caching, tiering or block level striping or the ability to serve storage off the network. The limitation of RAID cards simply seems to be software size.

      High end cards can do SSD caching, but AFAIK that's about the end of it past RAID sets.

      On the plus side for RAID cards is battery-backed write caches and freeing the OS from a certain level of overhead -- you can just unload your write on the card in very few cycles and let it deal with sorting out the parity and actually managing disk writes.

      You also get the advantage of a redundant OS boot LUN. I also find that software RAID tends to require more disks in the end -- you can have double parity with 4x disks that includes your OS and data.

      The sweet spot is probably RAID cards that support a subset of disks for redundancy and the rest in JBOD for OS-defined storage environments. We've done a couple of MS Storage Spaces installs -- 2x SATA hardware mirrored for boot, 2x SSD and 6X SATA in a tiered space.

    8. Re:Focus by Anonymous Coward · · Score: 0

      ZFS is high performance, better sequential and random IO than pretty much anything out there, and it scales. Make another FS really really large and see what happens, all kinds of random issues become a problem. Gotta FSCK? That's going to take a while on 1PiB of data. And why are you using a RAID controller anymore? ZFS has made those obsolete with better performance and better support.

    9. Re:Focus by arth1 · · Score: 1

      Not such a huge deal now since ZFS exists to do the RAID calculations.

      That doesn't help much if you have to go through a RAID controller, because that's the only way you can get a huge number of drives from particular vendors.

      ZFS is great, but a hardware controller will often support things like:

      - Boot partition is also on a RAID.

      - Ability to add a drive to a RAID, expanding it.

      - Global and local hot-spares.

      - No system load choke during rebuild. That's probably the #2 issue with zfs - systems are near unusably slow during rebuilds. Granted, the rebuilds take less time (at least if not near full), but during that time, it's not just write speed that's degraded like on a hardware RAID, but system load truly spikes.

      - Write-through with battery-backed up RAM. That's probably the #1 issue with zfs. You have to turn off all caching to get commits that can be trusted. Journaling helps preserve file system integrity, but does not guarantee that all individual transactions will survive also after a journal playback - unless all write caching is turned off. And then you kill zfs' advantage of being able to use gobs of system RAM for write caching.

      - Worst case predictability. With a quality hardware RAID, there's published speed data, including worst case times. With zfs, all bets are off. You can quickly get a decent measurement of the average speeds, but not the worst case write latencies. This is especially important when using SSDs, where worst case latencies per drive can be astonishingly high. Write amplification and whole sector re-initializations can give worst case writes in the 1+ second range, even if average write speed is super-fast. For some applications, it's not average speed that matters, but worst case.

      I'm a fan of ZFS, but it really doesn't do everything a hardware RAID does.
      Sure, you can still use a hardware RAID in BSD, but if you have to shut the system off to do maintenance because there is no RAID controller software that can run from userspace, it's not really a good choice for that.

    10. Re: Focus by dbIII · · Score: 1

      The reality is your disks are off the PCIe bus no matter what.

      The reality is that something has to do the RAID calculations and a single core 1GHz 32bit PowerPC is not as good at it as a Xeon with several cores each at least three times faster. You have to do those calcs before shovelling data to the disks, and if you can do it even before shoving it through the PCIe bus even better.
      With RAID6/raidz2 it's not going to take a lot before that single core 1GHz 32bit PowerPC can't keep the disks fed at full speed. It's a bottleneck.

    11. Re:Focus by dbIII · · Score: 1
      Several errors but here is the HUGE one.

      Sure, you can still use a hardware RAID in BSD, but if you have to shut the system off to do maintenance because there is no RAID controller software that can run from userspace

      There is "mfiutils" for the LSI stuff, even the older 3ware stuff has RAID controller software that can run from userspace on FreeBSD - via a web browser or via the 3dm2 tool. There is other stuff for other vendors RAID cards but I'm not familiar with it.

    12. Re: Focus by TheRaven64 · · Score: 1

      On the plus side for RAID cards is battery-backed write caches and freeing the OS from a certain level of overhead -- you can just unload your write on the card in very few cycles and let it deal with sorting out the parity and actually managing disk writes.

      This is not an issue for ZFS. RAID-Z doesn't have the RAID-5 write hole.

      You also get the advantage of a redundant OS boot LUN.

      FreeBSD can happily boot from a RAID-Z pool (though you do need to duplicate the bootloader on all disks, but that's under 1MB).

      --
      I am TheRaven on Soylent News
    13. Re: Focus by Billly+Gates · · Score: 1

      I will say a RAID card is ESSENTIAL for any type of database or financial processing or warehouse operation because of reliability and batteries in case of a failure.

      When the power goes off on a computer not all of the data can be swapped out to the disk on time.

      But for a home media server on Windows 7? Come on, 2001 was a long time ago :-)

      I would say these days with LUN people use eSAN for anything really important anyway and use just the CPU to boot the server. TRIM support is important for anyone using SSDs in the mix these days which lower end raid cards lack.

      --
      http://saveie6.com/
    14. Re: Focus by swb · · Score: 1

      Current SAS-12 LSI cards have a dual core 1.2Ghz CPU, which leads me to believe the bottleneck may have been eased and for conventional in-server applications probably isn't a real bottleneck.

      Mind you, I'm not disagreeing that there's more value in software raid, especially as disk counts go high, but for many in-server storage setups, especially spinning rust, you're going to exhaust disk throughput way before you exhaust the controller throughput.

    15. Re: Focus by dbIII · · Score: 1

      which leads me to believe the bottleneck may have been eased

      So long as you don't have to replace a failed disk and do a rebuild of the array to repopulate that disk.
      In that situation beyond a small array it's going to want as much CPU as it can get and not be bound by the disk speed.

    16. Re: Focus by dbIII · · Score: 1

      It would be essential if you don't have a file system designed to cope with pulling the plug. However there are several that are.
      I have not bothered to replace the batteries in my older RAID cards because there is no longer any need, just as it's now better to run them in JBOD mode.
      A lot of those SAN devices use ZFS and do the heavy lifting in software.

    17. Re: Focus by swb · · Score: 1

      But any array rebuild is inherently bound by the speed of a replaced disk as the remaining members can provide the data needed to reconstitute the replaced disk far faster than the disk itself can write it. The performance of the card in calculating parity info to do this hasn't been a significant issue that I can remember since the days of 5 x 1GB RAID 5 arrays some 20 years ago, in fact most cards have options to reduce rebuild rates below 100% to reduce performance drags on ongoing access. The drag isn't so much CPU bound but tied to random access thrashing on rotational media as rebuild reads aren't correlated with normal access patterns.

      With the advent of virtualization, nobody really builds out single server storage anyway -- it all migrated to FC or iSCSI SAN years ago. Where single-server RAID gets used is purpose built systems which combine some aspect of compute (backup, DB, etc) and there is a need for redundant storage independent of SAN.

      Although in-server storage is seeing something of a renaissance with the advent of hyper-converged storage, but this is all software and depending on the vendor often variable redundancy (ie, given LUNs can have specific redundancies, mirror, single or dual parity) with the added complexity of node level redundancy requirements (ie, what does double parity storage mean when you have single node redundancy?)

      I'm split on market uptake for this, as compute and storage scale differently and most hyperconverged solutions end up with excessive compute for needed storage and require compute and storage to be scaled in lockstep. I think compute platforms will need to get thinner (single CPU x multiple disk) to scale storage without the compute side making it absurdly costly (hardware, licensing, switching) when storage expansion is what's called for.

      But we're now seeing the cost of flash and 10/40gbe getting cheap enough where these distributed storage nodes can actually work and still deliver high quality I/O across distributed nodes.

    18. Re:Focus by epine · · Score: 1

      I'm late to this party, but just so anyone who stumbles upon this thread by some quirk of Google future, the views expressed above are not reliable. It's not apparent that the author knows much of anything about the ZIL or the SLOG. There are trade-offs involved with ZFS, no question. But none of these are anywhere as inane as this post would seem to have it.

      If the vast majority of your work load is synchronous write, you do have to provide a SLOG with as much write bandwidth as the rest of your pool. Except during recovery, the SLOG is pretty much sequential write-only (not a demanding case for any enterprise-grade write-optimized storage device). These writes take place concurrently to the synchronous writes (latency matters). Under ZFS, the primary pool still batches writes into transaction groups every few seconds. The comment about the inability of ZFS to use RAM for write caching is simply incoherent.

      What ZFS can't do, for synchronous writes, is use RAM for write coalescing, eliminating writes to stable storage (the SLOG, in this scenario) if one write immediately replaces another (a fairly common traffic pattern). Well duh you can only do this if your RAM counts as stable storage for any file system, and if you even have this, it's usually a device requiring I/O traffic to access, the same as any other persistent storage device.

      What hardware RAID does potentially buy you is combining both the persistent RAM and the persistent storage onto a single device channel, allowing the OS to kill two birds with a single I/O write operation. And for this, you buy yourself a really really complex layer of extra device firmware, which historically has been far from entirely bug free. Your surface area of failure increases enormously (though you do have fingers to point at the extremely well healed—all that internal firmware testing is baked into the price with a healthy insurance multiple—should the worst come to pass).

      Do you really need synchronous write coalescing? A basic Xeon these days has 40 lanes of PCIe 3.0. Does that look like a rate-limiting resource on sustained synchronous write traffic to your storage pool? I wish. And if it does, I'm pretty sure your first response is this: more sockets, please.

      As it happens, there are giant industry plans afoot to add a non-volatile memory type into the system memory hierarchy. ZFS will like this—a lot—should any of this chortling evil land-grab vapour come to pass.

    19. Re: Focus by dbIII · · Score: 1
      For what it's worth I've seen plenty of "resilvers" take around a quarter of the time an array rebuild with a replaced drive used to take on that same hardware before it had ZFS and used the CPU instead.
      Maybe those LSI RAID cards are crap compared with others but they definitely choke on doing the calculations.

      nobody really builds out single server storage anyway

      Apart from the people who do. There seems to be a lot of those "nobodies" in scientific computing. There seem to be a lot in other areas too.

  5. New and Improved!!! by NoNonAlphaCharsHere · · Score: -1, Flamebait

    Now with systemd!!!

    1. Re:New and Improved!!! by Anonymous Coward · · Score: 0

      Not at all!

    2. Re: New and Improved!!! by Anonymous Coward · · Score: 0

      The systemd developers have better things to do than support a dead OS.

      *BSD is dead

    3. Re:New and Improved!!! by Anonymous Coward · · Score: 0

      You wish, Lennart...

    4. Re: New and Improved!!! by Anonymous Coward · · Score: 0

      Does Netcraft confirm it?

    5. Re: New and Improved!!! by Hognoxious · · Score: 4, Interesting

      The systemd developers have better things to do than support a dead OS.

      Like killing a live one?

      --
      Confucius say, "Find worm in apple - bad. Find half a worm - worse."
  6. They just now added 802.11n support? by drinkypoo · · Score: 2

    Tell me that there's actually been a way to do it all along, but now there's just a better way.

    --
    "You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
    1. Re: They just now added 802.11n support? by Anonymous Coward · · Score: 0, Troll

      You're surprised that a dead OS is way behind on supporting wireless standards? Why?

      *BSD is dead.

    2. Re:They just now added 802.11n support? by Anonymous Coward · · Score: 0

      Next up: Support for that new-fangled USB 2.0 standard!

    3. Re: They just now added 802.11n support? by QuietLagoon · · Score: 4, Informative

      ...You're surprised that a dead OS... At peak times, Netflix generates over 35% of all the traffic on the US portion of the Internet. Guess what "dead" OS powers the servers that create all that traffic?

    4. Re: They just now added 802.11n support? by UnknownSoldier · · Score: 1

      How is *BSD dead when it runs the PS3 and PS4 ???

    5. Re: They just now added 802.11n support? by sconeu · · Score: 1

      I won't believe it's dead until Netcraft confirms it.

      --
      General Relativity: Space-time tells matter where to go; Matter tells space-time what shape to be.
    6. Re:They just now added 802.11n support? by Anonymous Coward · · Score: 1

      The FreeBSD slogan/motto/something is "the power to serve". Wireless standards just aren't a high priority for server operating systems.

    7. Re:They just now added 802.11n support? by Anonymous Coward · · Score: 0

      I too was susprised when I realized that servers tend to use cables for networking. Hilarious, no?

    8. Re: They just now added 802.11n support? by ls671 · · Score: 2

      How is *BSD dead when it runs the PS3 and PS4 ???

      Pretty amazing! I don't have PS3 on linux...

      ~/tmp$ set | grep PS
      GROUPS=()
      PS1='\[\e]0;\u@\h: \w\a\]${debian_chroot:+($debian_chroot)}\u@\h:\w\$ '
      PS2='> '
      PS4='+ '
      ~/tmp$

      --
      Everything I write is lies, read between the lines.
    9. Re: They just now added 802.11n support? by UnknownSoldier · · Score: 1

      I know you're being funny, but for those wondering about the context ...

      Sony uses *BSD for the PlayStation 3 and PlayStation 4 kernel.

    10. Re:They just now added 802.11n support? by drinkypoo · · Score: 3, Interesting

      I too was susprised when I realized that servers tend to use cables for networking. Hilarious, no?

      What's hilarious is all of these childish excuses.

      --
      "You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
    11. Re: They just now added 802.11n support? by Anonymous Coward · · Score: 0

      Because BSD is only good for hosting static files. Good luck trying to get any modern web dev language running on it without attempting to compile it yourself, and get frustrated with the dependency hell it creates.

      That said, still better than Windows.

    12. Re: They just now added 802.11n support? by Anonymous Coward · · Score: 0

      Obviously you haven't worked much with FreeBSD. But then maybe "pkg install python35" is too hard for you.

    13. Re:They just now added 802.11n support? by unixisc · · Score: 1

      What I'm wondering is whether they support WiFi ootb for Intel WiFi chipsets. I have a Dell Inspiron I7, and PC-BSD was unable to recognize the WiFi. Hopefully, 'TrueOS 11.0' would not have that shortcoming - although I'd have to get that DVD to do a new install

    14. Re:They just now added 802.11n support? by unixisc · · Score: 1

      In which case, they should have a 'downstream' team of PC-BSD/TrueOS driver writers who make sure that the OS works w/ laptops and AIOs that use WiFi and other desktop peripherals, such as printers.

    15. Re:They just now added 802.11n support? by Anonymous Coward · · Score: 0

      Tell me that there's actually been a way to do it all along, but now there's just a better way.

      It's been available since at least Dec 2011 in a development form.
      http://www.osnews.com/story/25461/Interview_with_Adrian_Chadd_on_802_11n_in_FreeBSD

    16. Re: They just now added 802.11n support? by ls671 · · Score: 1

      What is the "PlayStation" ?

      Is it an advanced Walkman ?

      Does it also have video capabilities?

      --
      Everything I write is lies, read between the lines.
    17. Re: They just now added 802.11n support? by _merlin · · Score: 3, Interesting

      Trouble is, whatever I get with pkg never has the options I need enabled, so I have to go back to using ports, then I get messy dependency issues "X needs Y to be newer than version B, but Z needs Y to be older than version A". With RHEL and similar, the binary packages tend to have kitchen sink enabled by default, which is better suited for my use cases.

    18. Re: They just now added 802.11n support? by Anonymous Coward · · Score: 0

      It's more appropriate to say they customized FreeBSD for their own purposes for both the PS3 and PS4. They do make use of the userland, as well as the kernel. In particular, they appreciate the port of LLVM and Clang to FreeBSD. The graphics stack I imagine is all theirs, but the networking stack is pretty much all FreeBSD, for example.

    19. Re: They just now added 802.11n support? by QuietLagoon · · Score: 1

      ....I get messy dependency issues "X needs Y to be newer than version B, but Z needs Y to be older than version A".

      Yeah, I just ran into that when I tried to install two different packages on the Debian 8 install on my notebook. I even tried a fresh install of Debian 8, and I still had problems doing pkg installs.

      .
      I've never seen it on FreeBSD, though. Even when I'm compiling ports.

      To each his/her own, I guess....

    20. Re:They just now added 802.11n support? by jimtheowl · · Score: 1

      What is sad is all of these childish comments.

    21. Re: They just now added 802.11n support? by Anonymous Coward · · Score: 1

      Good luck trying to get any modern web dev language running on it

      it's not all that complicated. pkg install .

      without attempting to compile it yourself, and get frustrated with the dependency hell it creates.

      What dependency hell? Update the Ports collection, and let the Ports system handle your dependencies. It doesn't get much simpler.

    22. Re: They just now added 802.11n support? by Anonymous Coward · · Score: 0

      pfsense seems to be doing well and my two firewalls I manage update regularly. They are running FreeBSD 10.3-RELEASE-p9
        now.

    23. Re: They just now added 802.11n support? by Anonymous Coward · · Score: 0

      No wifi. Less space than a Nomad. Lame.

    24. Re: They just now added 802.11n support? by Billly+Gates · · Score: 1

      Nope a mistake. They meant to say 802.11n support added for additional WiFi drivers in their errata.

      This reminds me of Kernel 2.6.0 mentioning supporting SMP mistakenly by the media.

      Boy the mcses at work were laughing and mention Server 2003 rules and had that for years! Grrr

      --
      http://saveie6.com/
    25. Re:They just now added 802.11n support? by adri · · Score: 1

      No, it's been in there since freebsd-9. i just taught more things about 11n and bugfixed what was there.

    26. Re: They just now added 802.11n support? by Anonymous Coward · · Score: 0

      On the kind of scale Netflix operate at, it wouldn't be too hard to have staging/prod clusters and centralised deployment of precompiled packages. I would be very surprised if this wasn't the way they do things.

      A little more difficult to prep than Linux, perhaps, but they obviously have people who know how to wrangle a BSD.

    27. Re:They just now added 802.11n support? by avgapon · · Score: 0

      In fact, FreeBSD had solid USB *3* support before Linux did.

    28. Re: They just now added 802.11n support? by drinkypoo · · Score: 1

      Nope a mistake. They meant to say 802.11n support added for additional WiFi drivers in their errata.

      Thanks for that. I was sure that this was not the whole story. *BSD might be behind Linux in many things, but not that far behind.

      --
      "You're right," Fisheye says. "I should have set it on 'whip' or 'chop.'"
    29. Re:They just now added 802.11n support? by Anonymous Coward · · Score: 0

      802.11g should be enough for everyone

    30. Re:They just now added 802.11n support? by TheRaven64 · · Score: 1

      I was a bit surprised about that, as my FreeBSD 10 machine has had working 802.11n since 10.0 (support was in head for a long time, but Adrian doesn't like to do MFCs so it took a while to make it into a release).

      --
      I am TheRaven on Soylent News
    31. Re: They just now added 802.11n support? by Anonymous Coward · · Score: 0

      Really? Pretty please do try to install Postfix with PostgreSQL backend on any RHEL. I dare ya. Ya can't. Because RH and friends are very conservative.

      Debian and friends cheat by spawning a bunch of packages. postfix, postfix-pgsql, postfix-mysql, postfix-this, postfix-that...

      On FreeBSD it's super easy to build your own pkg repo with whatever configs you want, something no linux distro has this easy, not even Gentoo which is super complicated if you want to build and maintain packages for multiple machines.

    32. Re: They just now added 802.11n support? by dnaumov · · Score: 1

      It's more appropriate to say they customized FreeBSD for their own purposes for both the PS3 and PS4. They do make use of the userland, as well as the kernel. In particular, they appreciate the port of LLVM and Clang to FreeBSD. The graphics stack I imagine is all theirs, but the networking stack is pretty much all FreeBSD, for example.

      This begs the question: PS4 has supported 802.11n since launch (November 2013), how come FreeBSD is only getting 802.11n support 3 years later?

    33. Re: They just now added 802.11n support? by Anonymous Coward · · Score: 0

      That's an easy one: the license.

      BSD license isn't an "as you receive, so shall you leave" license. Sony is not required by the BSD license to return improvements to the code. (GPL does not allow distribution of software without release of source.)

      So the Sony devs' improvements (802.11n) never made it back to "BSD central."

      In contrast, when Apple built Safari off Gecko (think it was), they returned improvements to the KDE devs - presumably to Konqueror's betterment.

    34. Re: They just now added 802.11n support? by Anonymous Coward · · Score: 0

      You're a complete fucking idiot blathering stupid shit.
      Please leave the discussion for the grown ups and go back to your playdough.

    35. Re: They just now added 802.11n support? by Anonymous Coward · · Score: 0

      In contrast, when Apple built Safari off Gecko (think it was), they returned improvements to the KDE devs - presumably to Konqueror's betterment.

      It's KHTML, not Gecko.

  7. great by Anonymous Coward · · Score: 0

    but does systemd support it?

    1. Re:great by Anonymous Coward · · Score: 0

      Yes:

      $ systemctl --version
      systemd 229
      +PAM +AUDIT +SELINUX +IMA +APPARMOR +SMACK +SYSVINIT +UTMP +LIBCRYPTSETUP +GCRYPT +GNUTLS +ACL +XZ -LZ4 +SECCOMP +BLKID +ELFUTILS +KMOD -IDN

      $ systemd-run --scope --user tmux qemu-system-x86_64 -enable-kvm -cpu host -drive id=hd,file=/path/to/freebsd.image,format=raw,cache=none -device virtio-scsi-pci,id=scsi -device scsi-hd,drive=hd -net nic,model=virtio -m 2G -monitor none -name "FreeBSD VM"

  8. *BSD is dead by Anonymous Coward · · Score: -1

    Fact: *BSD is dead, so dead in fact that nobody has updated the venerable *BSD is dying post in over a decade.

    Netcraft now confirms: *BSD is dead

    Once a viable server OS, *BSD has lost so much market share that this is almost exclusively the domain of Linux. This is confirmed by recent Netcraft statistics showing that only a tiny fraction of a percent of servers run *BSD.

    Although Usenet is a ghost town, discussions about operating systems have moved to other forums. The once thriving BSD section on Slashdot only receives a couple of posts per year, indicating that discussion about *BSD and development have slowed to a trickle.

    Despite increased fragmentation, as FreeBSD has been forked by projects like DragonflyBSD, none have been successful. Even attempts to fork *BSD have largely ceased, as developers have moved over to Linux.

    All this evidence points to a grim picture of *BSD: it is a zombie, gutted by the loss of developers to other projects as *BSD collapses in disarray. It would be false to say that *BSD is dying. In fact, *BSD is dead.

    1. Re:*BSD is dead by Anonymous Coward · · Score: 0

      Number of devices in my house running some derivative of BSD - about 12. Number running Linux - 1.

      Yup its pretty dead alright.

  9. FreeBSD by Anonymous Coward · · Score: -1

    No buttfucking allowed. Thank you.

  10. Slashdot Sponsored Whitepaper by FunkSoulBrother · · Score: 2

    Anyone seeing these? Any adblock/noscript rules that defeat them?

    1. Re:Slashdot Sponsored Whitepaper by Anonymous Coward · · Score: 0

      Nope, never seen anything like that. What does it look like?

      Browsing AC with uBlock origin, ghostery and noscript (last one but only partly blocking).

    2. Re:Slashdot Sponsored Whitepaper by FunkSoulBrother · · Score: 1

      http://i.imgur.com/ErnoXw5.png

      just running uBlock origin and noscript at the moment, though I do run noscript with everything globally allowed, then blacklisting domains that bother me/seem irrelevant as needed.

      I know, not the most secure but I'm more into it for making stupid JS design go away than out of paranoia of a script that doesn't visibly annoy me. Maybe I'll try ghostery or play with my NoScript settings. Thanks for responding!

    3. Re:Slashdot Sponsored Whitepaper by Shane_Optima · · Score: 1

      I think I saw those that a while back. uBlock element picker worked just fine on it, I think, so two clicks and its gone.

      uBlock is more or less a replacement for ABP, Noscript and Request Policy, although I think there are a few areas that NS covers that it doesn't (like anti-XSS). GPL, lighter memory footprint than ABP and it comes configured with Easylist out of the box. Very easy to use once you understand what the boxes in the GUI represent and you can effortlessly switch and combine whitelisting and blacklisting approaches. I can't recommend it highly enough.

    4. Re:Slashdot Sponsored Whitepaper by Shane_Optima · · Score: 1

      uBlock Origin*

    5. Re:Slashdot Sponsored Whitepaper by Shane_Optima · · Score: 1

      Ah ok, so you are using uBlock Origins. I think the element picker feature (the eye dropper thing) took care of it for me.

      As far as laziness goes, my uBlock Origin philosophy is this: try to run in a whitelist only mode if I'm not pressed for time, but as soon as I get a website running properly I immediately go through and globally blacklist (or sometimes just locally blacklist) everything I haven't whitelisted.

      This extra step only takes a few seconds, and it means that if/when you get frustrated or pressed for time, you can simply switch over to permissive / blacklist only mode. New domains will of course be permitted by default, but since you've been setting up blacklists as you go most of the junk should still be blocked on websites you've already visited (and if you blacklisted those domains globally, they will also be blocked on any new site you visit.) This is the best of both worlds, allowing you to be as lazy as often as you want without throwing away the effort you went through on those days when you actually took the time to manually whitelist each domain. (And of course when I say "whitelist", of course I mean gray/"no-op". I almost never use green/always-permit whitelisting.)

      For maximum fine-grained control, you can add uMatrix (written by the same guy) or Noscript into the mix but other than a lack of cookie management, I find that uBlock is more than sufficient for a daily driver.

    6. Re:Slashdot Sponsored Whitepaper by Anonymous Coward · · Score: 0

      Block library.slashdotmedia.com/ and they go away.

    7. Re:Slashdot Sponsored Whitepaper by Anonymous Coward · · Score: 0

      Er, sorry. That used to work just a week ago. Now they're served from a cloudfront.net url...

      I'm mostly irritated by the geolocation AJAX call.

  11. KMS support? by Billly+Gates · · Score: 2

    I know Hyper-V support has been improved from 10.3 as Azure has that custom port that MS contributed back.

    But KMS/Quemu interests me as any 2016 IT professional uses virtualization and VMare Workstation is discontinued and in life support mode and sucks greatly.

    --
    http://saveie6.com/
    1. Re:KMS support? by Shane_Optima · · Score: 2

      If you don't have specific requirements necessitating another solution (a longer support window, perhaps), I really can't recommend Qubes highly enough. You can run conventional Xen HVMs alongside Qubes AppVMs which 1. are damned fast, 2. can utilize shared templates for disk space efficiency and easier updates, 3. have tools for quickly and securely sharing files or clipboard contents and 4. allow you to intelligently mix and manage color-coded windows in a single task bar. These features are available with Windows 7 as well (Windows 10 is still in development.) Yes, they're stuck using Fedora for the base for now (they'll move to something minimal and more secure down the road) but you can have BSD or arbitrary Linux distro guests and Dom0 doesn't have any network access so there's no reason to be paranoid about it.

      The fact that it's probably the most secure full-featured desktop distro ever made (particularly on machines that properly support vt-d) is largely a result of it being a powerful hypervisor. Some people seem to think that it being an ultra-secure OS must mean its crippled or cumbersome but nothing could be further from the truth; it's been a usability win across the board and has been my daily driver for over a year now. I do wish the GUI had better built-in snapshot functionality, but backup functionality works fine and there's nothing preventing you from using btrfs and/or LVM.

    2. Re: KMS support? by Billly+Gates · · Score: 2

      I am talking about KMS and Freebsd as a host :-)

      Yes part of KVM was ported to the FreeBSD kernel. For now stuck on Windows 10 with Hyper-V for games. KMS provided GPU and hardware pass thru

      --
      http://saveie6.com/
    3. Re: KMS support? by Shane_Optima · · Score: 2

      Ok, yeah GPU passthrough is still a no-go for Qubes right now. It sounded like you might've been in the market for a general workhorse hypervisor, but 3d gaming is the one area in which it's definitely lacking.

      On the other hand, I should protest that Qubes does make for a superior casual gaming and retro/oldschool gaming experience. There's something to be said for running 2d Steam games (ones that don't have Linux compatibility), Fallout 2 and other oldschool Windows games in their own window sitting in the task bar right next to your Linux apps, with no WINE involvement and no Windows desktop or secondary taskbar in sight.

      (If you can't get 'em running in Windows 7 you will need to go through a secondary XP/98/95 desktop, through.)

    4. Re: KMS support? by Billly+Gates · · Score: 1

      I may take a look? I want to leave Windows but am so dependent on it. I do not want to deal with wine and need nested nested virtualization for my mcse Hyper-V labs to learn clustering. Yeah laugh at me on last bit.

      VMware workstation which is slow on my PC is being discontinued leaving me with just Hyper-V on Windows 10 ... But hey MS contributed awesome Hyper-V guest from azure that made it to FreeBSD 11 :-)

      --
      http://saveie6.com/
    5. Re:KMS support? by TheRaven64 · · Score: 1

      I'm a bit confused by KMS/Qemu. Do you mean KVM or KMS? Kernel Mode Switching (KMS) has been in FreeBSD since 10.0 and the graphics drivers are now very close to parity with Linux. If you mean KVM, FreeBSD ships with bhyve, the BSD Hypervisor, which is a legacy-free Type II hypervisor. It happily runs Linux and FreeBSD as guests and will run Windows Vista or newer with a little bit of fiddling. VirtualBox also works well on FreeBSD, if you want something more desktop friendly and FreeBSD works as a Xen Domain 0 if you want something a bit more cloudy.

      --
      I am TheRaven on Soylent News
    6. Re:KMS support? by Anonymous Coward · · Score: 0

      Since when is VMware Workstation discontinued? Show me the proof.

    7. Re:KMS support? by Billly+Gates · · Score: 1

      Since when is VMware Workstation discontinued? Show me the proof.

      Here

      http://www.theregister.co.uk/2...

      It still is on life support with bug fixes but and maybe a few fiddling with updating some guests but it is no longer actively developed. After playing with a faster type 1 hypervisor and seeing how quick and light it was close to bare metal I got spoiled. VMWare Workstation and Virtualbox is sluggish and always gets the cpu fans going with many bugs with emulating all the hardare

      --
      http://saveie6.com/
    8. Re:KMS support? by Billly+Gates · · Score: 1

      I meant KMS. From what I see it is no longer actively developed and quite behind the Linux version (correct me if I am wrong).

      This is just for desktop use at home as a workstation. I need type 1 speed and guest support and embedded or nested virtualization to learn some labs with clustering with other solutions like Hyper-V and VMware ESX. Last, I want to run my steam games too with full hardware acceleration.

      Linus tech tips got me interested in this with Unraid which is a distro of Linux KMS/QEMU with real 7 gamers on 1 CPU with full GPU pass thru which looks awesome!

      --
      http://saveie6.com/
    9. Re:KMS support? by unixisc · · Score: 1

      For FreeBSD, they have a different VM solution called Bhyve (pronounced Beehive). That, and also, they have jails for Debian and Gentoo distros of Linux.

    10. Re: KMS support? by Shane_Optima · · Score: 1

      I switched over from Virtualbox (I only ever tried VMware player, which I found pretty underwhelming), so I can't provide you with any useful comparisons. It's all Xen under the hood, if you've any experience with that.

      AppVMs, which use PV drivers (maybe PVH? I forget) and various performance tricks, are stupid fast to the point where you'll regularly forget that they're not native. Or maybe ~5 second boot times are the norm on most platforms now, but they certainly weren't on Virtualbox (even with PV drivers in use.) I think Windows HVMs were about the same speed as they were on Virtualbox, but the Qubes windows mixing (not having to deal with a second desktop) was a really nice touch.

      The point of the Qubes distro, it's worth underlining, is the intelligent plumbing, UIs, and security. The actual nuts and bolts of virtualization is left to the hypervisor, which as I said is Xen. (They're working on abstracting everything so that other virtualization platforms can be used down the road, but it isn't user-changeable at this time.)

    11. Re:KMS support? by Anonymous Coward · · Score: 0

      That's not an announcement of an EOL of a product, it's speculation based on HR practices (You do know about EMC being bought out by Dell, right?), I think I'll wait for a real announcement before relying on The Register.

  12. Just Released? by Anonymous Coward · · Score: 0

    Just released? 11.0-RELEASE has been on their FTP for about 2 weeks and i've been running it about that long now too =) 11.0-RELEASE-p1 is already available by freebsd-update

    1. Re:Just Released? by Anonymous Coward · · Score: 0

      It hasn't. A file named 11.0-RELEASE has been on their FTP site for week, but the actual release was not until just now.

  13. gimme a pitch on FreeBSD by AnAlchemist · · Score: 2

    I've been a long-time Linux user, but I'm not religious about it, and I've always been curious about the BSDs.

    Can someone give me an elevator pitch, especially about FreeBSD, seemingly the most popular of the BSDs? All the (server) software I use on a regular basis runs on FreeBSD.

    Before someone says "just try it," there's sooo much cool stuff to try (currently learning Clojure and Raspberry Pi stuff), so I need a reason to try it.

    Gimme some.

    1. Re:gimme a pitch on FreeBSD by otaku244 · · Score: 1

      Personally, I am a big fan of Jails over VMs. It's a much smaller footprint and essentially has as good of a sandbox. I haven't run it in a production environment, but by logically breaking out my jails into more discreet functions like I would with VMs (in my home environment), I feel like I have excellent control and reasonable security I would otherwise not have if all the apps were on the same platform.

      --
      Mod me down, I shall become more off-topic than you could possibly imagine.
    2. Re:gimme a pitch on FreeBSD by Anonymous Coward · · Score: 1

      It does what linux does just slightly more conveniently and elegantly, fewer problems at the margins. ZFS on freebsd has somewhat fewer problems than on linux, its also better integrated with other os pieces. Jails on freebsd are easier to use and work "better" (in my experience and in somewhat limited cases) than lxc containers. It will allegedly do high volume service faster than linux (haven't tried it have no need) and it is MUCH better documented than linux. I think for someone with motivation to leave the os x / windows ecosystem and start using an *nix freebsd is easier to get started with due to the great handbook they have.

      That said, to be fair, freebsd really shines as a server setup, it (in my opinion) fails pretty badly as a desktop os. Some of the linux distros, like mint, make phenomenally awesome desktops. All of the *nix's still have problems here and there with laptops though. I still think windows or os x are a better choice as a laptop computing os.

    3. Re:gimme a pitch on FreeBSD by Neo-Rio-101 · · Score: 2

      FreeBSD has been traditionally used in ISPs, primarily for it's networking.
      It also supports native ZFS for storage

      It's usually been the base OS for other projects like pfsense, and FreeNAS

      Also given it's licence, any changes made to the source don't have to be shared... so code gets creatively "borrowed" from it a lot.
      Just ask Apple.

      --
      READY.
      PRINT ""+-0
    4. Re:gimme a pitch on FreeBSD by dbIII · · Score: 2

      Can someone give me an elevator pitch

      If you want something like ZFS it's a couple of years ahead of linux.
      Otherwise it's fairly similar and slightly better ft on older hardware than current versions of linux.
      Not much point in trying it without one of those two reasons since it's really very similar to what you are used to.

      As a final note, the ports collection on FreeBSD appears to be the Gentoo linux dream achieved. Just tick boxes instead of choosing compile flags.

    5. Re:gimme a pitch on FreeBSD by Anonymous Coward · · Score: 0

      As a final note, the ports collection on FreeBSD appears to be the Gentoo linux dream achieved. Just tick boxes instead of choosing compile flags.

      Which is deeply amusing, given that FreeBSD inspired Gentoo in the first place.
      Granted, it's been many, many years since I've used Gentoo, but when I did, it was basically kinda like FBSD except global update constantly broke everything, a huge chung of ebuilds didn't actually compile, and there were never binary packages available.

    6. Re:gimme a pitch on FreeBSD by TheRaven64 · · Score: 1

      As a final note, the ports collection on FreeBSD appears to be the Gentoo linux dream achieved. Just tick boxes instead of choosing compile flags.

      Note that, if you want to compile ports yourself, it's now *strongly* recommended to use Poudriere rather than compiling individual ports stand-alone. Poudriere can compile any subset of the ports tree you want and give you a consistent package set. Poudriere builds ports in a jail so they won't ever accidentally be affected by other things on your system and will always only have the dependencies that are explicitly set. You can then install the packages alongside the upstream ones or, if you want to do something more interesting (e.g. building everything with a different compiler) build just the ones that you want.

      --
      I am TheRaven on Soylent News
    7. Re:gimme a pitch on FreeBSD by unixisc · · Score: 1

      One thing - does v11 now have SteamOS jails? This was supposed to be there - previous versions had jails for Debian and Gentoo (but not Fedora)

    8. Re:gimme a pitch on FreeBSD by shutdown+-p+now · · Score: 1

      Any specific answer would be very point-in-time (hardware and software support etc changes).

      Philosophy-wise, though, there is a difference that is a constant. In FreeBSD, the people who write the kernel are the same people who write libc and the rest of the base system (or at least package and test it). So the core OS feels less like something assembled with glue and duct tape, and more like a single polished product. It's even more obvious when you look at the source - same coding standards everywhere, same quality bars etc.

      Whether it's a good thing or a bad thing or irrelevant depends on your perspective.

  14. Death to syndromeD by Anonymous Coward · · Score: 0

    Title.

  15. coreutils side by side by itomato · · Score: 1
    Kernel and Network stacks aside, I have found (some) enlightenment comparing BSD and GNU coreutils:

    cat, for example:
    • http://git.savannah.gnu.org/cgit/coreutils.git/plain/src/cat.c
    • https://github.com/freebsd/freebsd/blob/master/bin/cat/cat.c

    ls is interesting:

    • http://git.savannah.gnu.org/cgit/coreutils.git/plain/src/ls.c
    • https://github.com/freebsd/freebsd/blob/master/bin/cat/cat.c

    Then again, this comparison has also driven me closer to suckless.

    • http://git.suckless.org/sbase/tree/cat.c
    • http://git.suckless.org/sbase/tree/ls.c
    1. Re:coreutils side by side by Anonymous Coward · · Score: 0

      ls is interesting:

      • http://git.savannah.gnu.org/cgit/coreutils.git/plain/src/ls.c
      • https://github.com/freebsd/freebsd/blob/master/bin/cat/cat.c

      ls is indeed interesting. FreeBSD's ls(1) looks just like cat(1)!

    2. Re: coreutils side by side by itomato · · Score: 1

      Opps!

  16. They should? by Anonymous Coward · · Score: 0

    Yes, they should have a cornucopia of downstream teams. Will you pay for it? No?

    1. Re:They should? by unixisc · · Score: 1

      If they have a separate PC-BSD/TrueOS project, why not have within their mission coverage objectives of not just making it prettier and easier UI (Lumina, PBI, AppCafe, et al) but also all device drivers that are irrelevant for servers (like WiFi) but pretty relevant for desktop distros based on it, like TrueOS?

  17. I'm a bit confused by Anonymous Coward · · Score: 0

    DriverS? Plural? FreeBSD 11 only seems to support Intel GPUs up to Haswell, old Radeon drivers, has no AMDGPU support for modern AMD GPUs, no Nouveau and none of the other random assortment of video card drivers that Linux has.

  18. Wrong on ZFS by Anonymous Coward · · Score: 0

    You are wrong on several things on ZFS. First of all, ZFS protects your data against data corruption. Hardware raid does not do that. Your data might be corrupt. Several independent researchers show that ZFS are safe.
    https://en.wikipedia.org/wiki/ZFS#Data_integrity
    Hardware raid is not safe:
    https://en.wikipedia.org/wiki/RAID#Weaknesses

    Second, ZFS scales across many HBA cards on a single server. You can connect many 100s of disks to several HBA disk cards. This means ZFS scales to thousands of discs on a single server. Hardware raid does not, it scales to a few 10s of disks. There are many Petabyte ZFS raids out there reaching extremely high throughput and performance. You will never see such large hardware raids.

    Clustered Lustre filesystem once used ext3 as a backing store, to save all the data on the disks. But ext3 had performance issues and scaling problems so they switched to ZFS. IBM Sequioa supercomputer uses a 55 Petabyte Lustre installation using ZFS, with 1 Terabyte/sec bandwidth. Try that on a hw raid. ZFS has much higher perforamnce than a raid card, because ZFS scales across many HBA cards.

    Third, Oracle ZFS has rewritten resilver mechanism and now ZFS resilvers (repairs) a disk at full platter speed, close to 150 MB/sec. It is not slow, you say ZFS resilver is slow with large spikes. Well, it is very very fast, much faster than hw raid.
    http://milek.blogspot.se/2014/12/zfs-raid-z-resilvering.html

  19. TCP IP stack superior by Anonymous Coward · · Score: 0

    FreeBSD have a superior TCP/IP stack. Much research are carried out on FreeBSD, not Linux.
    http://www.theinquirer.net/inquirer/news/2359272/facebook-wants-linux-network-stack-to-rival-or-exceed-freebsd
    Facebook seeks a developer to improve Linux stack:
    "...the job description begins, "Facebook is seeking a Linux Kernel Software Engineer to join our Kernel team, with a primary focus on the networking subsystem. Our goal over the next few years is for the Linux kernel network stack to rival or exceed that of FreeBSD." It concludes, "This position is full-time and located in our Menlo Park office."..."

  20. ZFS safe, hw raid not by Anonymous Coward · · Score: 0

    Research shows that ZFS is the safest storage solution out there. Hardware raid is not safe and might corrupt your data. ZFS protects your data, according to comp sci researchers:
    https://en.wikipedia.org/wiki/ZFS#Data_integrity
    HW raid cards are unsafe and might corrupt your data:
    https://en.wikipedia.org/wiki/RAID#Weaknesses

    And besides, ZFS scales over several HBA cards on a single server. So you can span a large ZFS raid across hundreds (thousands?) of disks on a single server. There are large Petabyte ZFS raids out there. HW raid cards scale to 10s of disks. When you have many 100s of disks you reach very high performance. ZFS is therefore much much faster than a single hw raid card. Just google on large petabyte zfs installations. And compare them to a single hw raid card. HW raids are not safe either, and slower. And cost money. ZFS is free, safe, and faster.