FreeNAS Switching From FreeBSD To Debian Linux

dnaumov writes "FreeNAS, a popular, free NAS solution, is moving away from using FreeBSD as its underlying core OS and switching to Debian Linux. Version 0.8 of FreeNAS as well as all further releases are going to be based on Linux, while the FreeBSD-based 0.7 branch of FreeNAS is going into maintenance-only mode, according to main developer Volker Theile. A discussion about the switch, including comments from the developers, can be found on the FreeNAS SourceForge discussion forum. Some users applaud the change, which promises improved hardware compatibility, while others voice concerns regarding the future of their existing setups and lack of ZFS support in Linux."

Comments

Hmmm

[Tellarin]

I don't want to be inflammatory, but having "concerns regarding the future of their existing setups" when using a piece of free software in version 0.7 for something as important as data storage?

I'd say that if your setup is so important you care so much for its future and you're facing this scenario, you have bigger concerns to take care than a move from FreeBSD to Linux.

Re:Hmmm

[Enleth]

You are not inflammatory, you just give more meaning to the position of the first decimal point in the version number than it deserves.

Would the software magically be better if the version was 8.0? 2009.12? 3.141592? 666.123.789? There are many post-1.0 applications that are hopeless, buggy crap, quite a bit of them even commercial, and just as much sub-1.0 software of high stability and overall quality.

In this case, as with many FOSS projects, the sub-1.0 numbers probably mean "there are still features to be added before we consider our work complete". The keywords are "we", "consider" and "complete". "We" != "any other user with a different set of requirements", "consider" != "claim as absolute truth", "complete" != "stable". In other words, a 0.8 version might be perfectly stable, just not feature-complete from the author's point of view, and perfetly sufficient for a subset of potential users with less sophisticated needs.

And why 0.8 and not 2.3.075? My best guess is "because they could and they liked it better."

Case closed, have a good day.

Re:Hmmm

[Tellarin]

In this case, as with many FOSS projects, the sub-1.0 numbers probably mean "there are still features to be added before we consider our work complete".

I'd change your definition to "before we consider the initial version of our work complete". This is exactly why I mentioned sub 1.0 version number in a piece of free software. It means there is no marketing department requiring bumping up the version number to impress anybody.

So, as you say, the devs themselves don't think it has the capabilities to be granted the 1.0 number. For whatever reasons they feel.

In other words, a 0.8 version might be perfectly stable, just not feature-complete from the author's point of view, and perfetly sufficient for a subset of potential users with less sophisticated needs.

The key word here is "might". It might, it might not. One also has to consider that even if the system does have all the features you want and seems stable, is it being properly tested and maintained? Has it been around long enough for it to count as some indication that the devs aren't going to just give up on it soon? Is there already a community around it?

All of this goes into choosing a sub-1.0 project for something important. This is what I meant. To depend on an early version of a piece of software is too big of a commitment without the proper analysis of these and many other issues, most of which are not related to the features per se.

And, in any case, it is free software. So anyone can fork the project and continue with it. And it seems there is actually a fork of this project to keep it running over FreeBSD.

None of this changes what I said. If whoever is using it and worried about its future did consider this issues, good for them. If not, well...

Re:Hmmm

[zhilla2]

Yes but this 0.7 version is just a repacked / modded FreeBSD 7.2, OS that's been in development since 1993 - and that itself was a fork of older projects. Much better than software embedded some hastily released commercial NAS.

Re:Hmmm

[OrangeTide]

If a vendor isn't willing to go to 1.0 then why should a customer have confidence? 1.0 is a milestone. Certainly it has absolutely no technical meaning, but that does not mean it has no meaning at all.

Re:Hmmm

[Anonymous Coward]

FreeNAS is an "easy-to-use" NAS for old hardware, and light on documentation -- read: it has a wiki; generate your own. So it's going to get a lot of first-timers, however technical, and they're going to have questions about the migration. Hence "concerns" in this sense really shouldn't be read as 'emotional outbursts of near panic', but as inquiries.

Anyhow, the traditional /. missing link for this story would be: http://www.learnfreenas.com/blog/

...Today Olivier Cochard-Labbé has made a great announcement, FreeNAS will live on and production ready ZFS support will be added with the upgrade to FreeBSD 8.0. At the same time a new Linux version of FreeNAS will be created called OpenMediaVault! Olivier explained it like this: FreeNAS needs some big modification to remove its present limitations (with one of the biggest being the lack of support for add-ons/plugins). We think that a full-rewrite of the FreeNAS base is needed. Therefore, we will take 2 different paths:...

I guess /. is running the story because it's a migration from a BSD to a Linux. But it's a nice minor news items on an interesting project, and is mostly useful by bringing FreeNAS to the attention of /.'ers who are starting to think about setting up a NAS.

Re:Hmmm

[RocketRabbit]

Version a default installation of any free Unix clone, and you will see that many, many apps are at less than 0.7.

Version numbers are totally meaningless.

Re:Hmmm

[Antique Geekmeister]

Or, they don't care. Changes in major release number often mean incompatible features. I'd have given a lot, for example, for OpenSSL to use a sane numbering system and release "0.97" as "9.7", and "0.98" as "9.8". Or the idiots over at CPAN who release version 1.1, 1.2, 1.21, 2.2105, then 1.3.

Re:Hmmm

[nine-times]

Well there are a couple things to consider. First, as others have mentioned, the fact that it's not yet at 1.0 has less importance in many free software projects. There are definitely projects that haven't reached 1.0 yet are still production ready.

Also, you seem to throw "free" in there like it's a bad thing (or maybe I'm misreading?), but the fact that people are using an open-source project may indicate that they're very concerned about future maintenance of the software they're using. Being open source prevents the possibility of it simply being dropped, leaving the users no recourse. Free software can be passed into others' hands or forked, and in the worst case scenario a business can theoretically pay a programmer to fix problems.

So in that sense, at least, free software is pretty good for future-proofing. And even if the whole setup is not yet at 1.0, many of the components are quite robust and mature. Both FreeBSD and Debian are very good, as are Samba, ProFTP, and the other services.

Re:Hmmm

[Enleth]

The key word here is "might". It might, it might not.

And the very same thing can be said about 8.0, 3.141592 or 666.456.789 with the same implications about being properly tested and maintained and other things you pointed out in the rest of the quoted paragraph. I've seen too much post-1.0 buggy crap to believe otherwise.

I'd change your definition to "before we consider the initial version of our work complete". This is exactly why I mentioned sub 1.0 version number in a piece of free software. It means there is no marketing department requiring bumping up the version number to impress anybody.

So, as you say, the devs themselves don't think it has the capabilities to be granted the 1.0 number. For whatever reasons they feel.

Well, that's what I had in mind. 1.0 is the point where the devs can do a high fiver, pull out the bottle of champagne and (when sober again) start thinking about completely new features. The exact point of completeness is, however, absolutely arbitrary. There are projects being released under version numbers in the 0.1-0.5 range with more features and better stability than similar projects already at 2.x or something like that, because the latter were aimed for a much smaller feature set. Interestingly, there are even projects that go for 1.0 asymptotically - the philosophy behind them being that 1.0 is a perfect solution to a given problem (as opposed to your "initial version" view), and perfect is impossible, so an actual implementation should never be released as "1.0". Still other projects are using the 0.1-1.0 range much like 1-10, to indicate further *stable* generations (the minor version number then belongs after the second decimal point), and only advance the first number on major project philosophy and structure changes (like, switching programming languages or major libraries, redoing big parts from scratch for better maintainability, turning a client-side GUI app into a web app, adding a whole completely new and different set of tools or features, changing protocols or file formats without backwards-compatibility, etc.), which might actually result in the version 1.0.15 being much less stable and complete than, say, 0.7.46.

All in all, making a broad statement about a project based primarily on its version number is IMHO more in the realm of haphazard numerology than engineering and professional (as in "done in accordance with the best practices", not necessarily "done for money") risk assessment for software deployment.

Re:Hmmm

[A nonymous Coward]

Version numbers are totally meaningless.

No, they are a significant milestone in how people think of a project. You can't compare versions of different packages, but version 1.0 means *someone* thought it significant, whether that be marketing, devs, or the guy running the ouija board. For a free source unpaid project, it means the devs think things have stabilized significantly. It has no significance in relation to other projects or indeed any one not heavily involved with the project.

Re:Hmmm

[laird]

"I don't want to be inflammatory, but having "concerns regarding the future of their existing setups" when using a piece of free software in version 0.7 for something as important as data storage?"

You are kidding, right?

First, the meaning of version numbers is purely dependent on the project. There are plenty of pieces of software that were fine to use in production with version numbers less than 1.0, and there are plenty of pieces of software with larger version numbers (e.g. every other Linux kernel version) that should not be run in production. FreeNAS extremely stable.

Second, even if the version of the FreeNAS that people used was not stable, it is based on a solid OS, filesystem, file services, web server, etc., that are all quite reliable in production, so even if FreeNAS 0.7 had a problem, the problem would be with some UI or scripting, not with the ability to store and serve data, and the data itself would not be at risk. The reality, of course, is that FreeNAS is one of the more mature, reliable NAS projects, so it doesn't matter much that you don't like the number 0.7.

The real problem here isn't that people were stupid to use FreeNAS, but that the FreeNAS developer is making a controversial move of switching the underlying OS to one that doesn't support a previously supported, and very popular, filesystem. And since data storage is the point of a NAS, that change is unpopular with the users of FreeNAS. The same change would have the same impact whether it was numbered 0.7, 1.0 or 3.1. For example, look at the upgrades from XP (AKA Windows 5) to Windows 7, which requires a data backup and restore, and reinstallation of all applications. I suppose you could attempt to argue that XP was too unstable to rely on, and that upgraders were stupid to have relied on such an immature OS, but that isn't what actually happened. Just as with FreeBSD, the developer chose to make a somewhat incompatible upgrade, and the users have to deal with the fallout.

From my perspective, my data, which is all in ZFS, is more valuable than the particular web/admin tools used to serve the data, so it means that FreeNAS 0.8 won't be an option for me. Luckily there are plenty of options, and migration is easy.

Re:Hmmm

[bill_mcgonigle]

but version 1.0 means *someone* thought it significant

Call me old fashioned, but I think the practice of numbering releases with ordinal numbers is far less confusing. Alpha and beta monikers mean 'still adding features' and 'working out bugs' respectively. One dot (1.1) releases are feature changes, two dot (1.1.1) are bug fixes, and no-dot (2.0) reflect significant re-works (architectural, interface, etc.).

0.87 tells me nothing based on the number. 4.1b7 encodes lots of information.

Re:Hmmm

[DamnStupidElf]

I'd rather have a project whose goal is "well tested and bug free" instead of "reaching milestones." There's always time to add some feature later, but no way to get your lost time back if things break.

Re:Hmmm

[evilviper]

If a vendor isn't willing to go to 1.0 then why should a customer have confidence?

Because they've done their testing of the software and found it to be highly reliable?
Because they've been following development, reading release notes, etc., and know what is and isn't stable, forward compatible, and/or feature complete?
Because they're not idiots who base their software decisions on intangible version numbers, rather than any actual knowledge of the product?

1.0 is a milestone.

0.7 is a milestone. 2.0.5 is a milestone. Any special significance 1.0 has TO YOU is just that, an invention of your own mind, irrelevant to the rest of the world, no more significant than rules such as avoiding odd-numbered Star Trek films, your favorite color, etc.

Certainly it has absolutely no technical meaning, but that does not mean it has no meaning at all.

Yes it does.

One buggy piece of junk may be version 9.7.5, while another highly reliable and usable piece of software may be version 0.0.1.

THERE IS NO PENALTY FOR NAMING YOUR BUGGY SOFTWARE VERSION-1.0, NOR ANY TANGIBLE BENEFIT TO INCREASING THE VERSION NUMBER ON STABLE SOFTWARE. THEREFORE, IT MEANS NOTHING AT ALL.

Re:Hmmm

[jonadab]

Version 1.0? Meh. Emacs is up to version 23.1. Heck, way back when I was still using Windows with Cygwin and I compiled my own Emacs for the very first time, it was version 21.0.105 way back then, and that was years ago.

Why should I use any software with a version number less than 20? Everyone should get with the times and upgrade to using Emacs for everything. If you aren't using Emacs to post to slashdot, your software is not up to date or stable enough. Demand better: switch to Emacs.

If that reasoning doesn't make sense, it could be because comparing version numbers from one piece of software to another is inherently meaningless.

(Note: I *do* use Emacs for a lot of stuff. But I also use other software sometimes.)

Re:Hmmm

[SydShamino]

Vendor != Developer

If a vendor wants to base their product on FreeNAS 0.8, they are more than welcome to call their product NASPoint 1.0 or NASPoint 2010 or whatever the hell they want.

The developers of FreeNAS use a number less than 1.0 because they feel their product is not yet feature complete for a 1.0 release. Obviously, one of the issues is a reliance on an underlying operating system which they see as hampering them in some way, so they are going to switch operating systems before 1.0.

Re:Hmmm

[mvdw]

Or that other software house that released 3.0, 3.10, 3.11, 95, NT, 98, 2000 (oops, version numbers not 2k compliant), XP, 2003, vista, 7. Make up your mind people. Version numbers, years or names. Pick one, please. Your inconsistency does not reflect well.

Re:Hmmm

[dbIII]

I'd change your definition to "before we consider the initial version of our work complete"

From the context it is extremely obvious that is definitely not the case. You are just grumbling because they are numbering releases differently to the way you would do it.

Re:Hmmm

[Anonymous Coward]

Actually, from 95 on, those are product names, not version numbers. Windows 7, for example, is really NT 6.1. Windows 2000 was NT 5.0, and Windows XP was 5.1 (I think... something like that anyway)

Much like Linux distributions have names (Ubuntu Karmic Koala, for example), but also version numbers (Karmic Koala is version 9.10).

Product name != version number, even if the name does happen to have a number in it.

Re:Hmmm

[OrangeTide]

bingo. I agree entirely.

Re:Hmmm

[Tellarin]

What part of me saying

the devs themselves don't think it has the capabilities to be granted the 1.0 number. For whatever reasons they feel.

makes you think I'm trying to guess how they number their releases?

Re:Hmmm

[ToasterMonkey]

0.7 is a milestone. 2.0.5 is a milestone. Any special significance 1.0 has TO YOU is just that, an invention of your own mind, irrelevant to the rest of the world,

Really? Let me get this straight, we're talking about revision numbering schemes, and you're saying that "1" has no special meaning? None at all? The rest of the world agrees?

Are you fucking nuts? Of course it has no technical meaning, it is a revision identifier. It is also the first number, genius. Meaning the first of something in most damned cultures. There may have been rough drafts, demos, tests, mock ups, etc with prerelease version tracking identifiers, but "1" is the one you put in front of your audience as the first release. How you can get up on /. of all places and say labeling a product "1" has no special meaning is insane. The next most meaningful identifier would be "2", arguably your first real version number.

You are free to throw any made up numbering system along with any unrefined piece of trash at your audience, but having the balls to stand up and tell them explicitly that THIS is my product, as-built, "one dot fucking oh" is extremely meaningful.

Re:Hmmm

[dbIII]

Please note that the "initial release" was some time ago which makes the text I replied to incorrect. That should show that your musings about version numbering systems are different to theirs and that your reply to me is missing the point.

Re:Hmmm

[RocketRabbit]

No, it doesn't mean anything. It should, that's something I would argue, but it doesn't.

Re:Hmmm

[BikeHelmet]

I believe Microsoft set the Win7 version number to 7.0, even though there were no significant kernel changes to warrant it.

I mean, Win2k to XP had kernel changes, but it was only .1 - not a complete overhaul. Same story for Vista -> Win7.

Your point about product names vs version numbers is totally correct.

Re:Hmmm

[BikeHelmet]

I hope they use an older kernel. The new linux kernels have great speed, and total shit compatibility. I've got four old PCs sitting next to me, and not one of them can detect a DVD drive or read stably off a PATA HDD in Ubuntu 9.10 - yet it works fine in 9.04 (minus the CD drives), and in 8.10/8.04 everything is working great.

It's sad to see an OS cut support for old hardware, like a 2.2ghz Athlon XP w/ 2GB of RAM running off a 120GB PATA HDD. (Still fast enough to browse the internet :P )

Re:Hmmm

[the real darkskye]

Microsoft Windows [Version 6.1.7600]

Copyright (c) 2009 Microsoft Corporation. All rights reserved.

It says 6.1 for me (7 x64 Pro).

Re:Hmmm

[Sxooter]

Sounds like you spend too much time reading marketing materials and not enough time testing things.

Re:Hmmm

[BikeHelmet]

Well that's good. :) An engineer somewhere got his way.

Re:Hmmm

[badkarmadayaccount]

Firefox is the new emacs. Please notice similarities between extension languages and feature creep, as well as developer/user in-jokes about memory/CPU usage.

Re:Hmmm

[badkarmadayaccount]

s/older/non-generic/g
s/new/new generic/g
You realize MFM drive support was rather recently removed from mainline, right?

Well, it's open source, so fork it.

[Anonymous Coward]

From the last page of comments, it looks like one company is already forking it to keep it on FreeBSD.

Half of the comments are users who picked FreeNAS for it's ZFS functionality worrying that they were stuck on FreeNAS 0.7.

Greater hardware compatibility? Sure, for some desktop computer hardware, but FreeBSD is fine for everything a NAS needs.

Re:Well, it's open source, so fork it.

[Anonymous Coward]

Not only that, but FreeBSD is a far more reliable and higher-quality core than even Debian could ever hope to be.

The FreeBSD development process and team is far more integrated and centralized. This has resulted in a codebase that is much cleaner than what we see in the more distributed development model non-BSD open source software (including Linux).

Changes and new features go through a strenuous review process before they're admitted to the FreeBSD codebase. If code makes it into a public release of FreeBSD, you can be damn sure that it is of an extremely high quality, and has been reviewed by some of the best minds in the field.

This isn't as much the case with Linux and much of the userland software that Debian uses. The quality of the code is generally lower than that of FreeBSD's code, and bugs can creep in much easier.

For something as critical as storage, FreeBSD is clearly the way to go.

Re:Well, it's open source, so fork it.

[Cyberax]

"Not only that, but FreeBSD is a far more reliable and higher-quality core than even Debian could ever hope to be."

[citation needed]

Linux is definitely faster and more feature-rich than FreeBSD. About the only advantage of FreeBSD is ZFS, and that's being fixed by btrfs.

Re:Well, it's open source, so fork it.

[dougisfunny]

I would imagine anything that makes it into Debian stable is pretty comparable, given the two years or so between Debian releases. Now if they're using Debian testing as a base, I would concede that its quite feasible that the code could be less stable than a FreeBSD based distro.

Re:Well, it's open source, so fork it.

[Virak]

Well I'm glad you're here to finally save the day and free all those big businesses relying heavily on Linux on their servers with no problems from their OS that apparently drops data like a quadriplegic juggler. Thanks, anonymous FreeBSD fanboy, the world would be a worse place without you.

Re:Well, it's open source, so fork it.

[Anonymous Coward]

Blah blah blah. If that actually was true, FreeBSD 5.x would never have been released. It wasn't nearly as robust as any release of Debian (granted, it was a remarkably poor release, especially for FreeBSD); and no release of FreeBSD, ever, is nearly as well tested as Debian. That's actual fact.

Your comment is nothing but fanboy idiocy, and is as "insightful" as a Coca-Cola advert.

Sure, FreeBSD is good, and in some ways better than any Linux distro. But generally better? No.

Re:Well, it's open source, so fork it.

[imp]

It all depends on what FreeNAS' target market is going to be. Is it going to be old desktop machines that people recycle into NAS boxes, or will it be the large variety of NAS boxes that are found in the wild today. If the former, then the switch to Linux buys you nothing. Really, FreeBSD and Linux run the same on x86 hardware (sometimes one is faster, or the other, or there's an issue that keeps one or the other from running, but in general both just work damn well). If the target is the latter, then Linux might have a small edge, but only because the FreeBSD project hasn't focused on the proper packaging of FreeBSD for an embedded system that has the tight memory constraints that the non-intel NAS boxes have. Many companies have climbed this hill, but there's nothing that's been standardized enough to be ready to include in FreeBSD (although both NanoBSD and TinyBSD could be made to work). M0m0wall and FreeNAS innovated in other areas, and this area would be easy to innovate in as well, since the problem is well understood and most of the tools necessary to make it work are already extant in the tree.

Forking FreeNAS may or may not be the right thing to do. It might be better to provide a FreeNAS 0.7 -> NewFreeNAS project that is rewritten from scratch for FreeBSD 8.0 that doesn't suffer from the php interface that replaces /etc/rc.d. That's the main barrier to porting from 7.x -> 8.x for FreeNAS (and m0m0wall). It would likely be faster and simpler to go that route and fix whatever issues come up. This would allow one to migrate to better http technology that puts less in the server and more on the client in javascript/ajaxish/etc things anyway. This would allow users to continue to use FreeBSD's solid ZFS base as well as have a solution that's here today rather than waiting for Linux to catch up with its reimplementation of zfs :)

Warner

Re:Well, it's open source, so fork it.

[Anonymous Coward]

Linux changes go through a review process by some of the best minds of the field (also called "maintainers"), in fact most of them are paid to work full time in Linux, which improves the quality.

One of the main problems that companies find today when contributing to Linux is that it's too difficult to contribute code to Linux because it requires too many effort to get the code merged. That's because Linux doesn't have a "unstable" development phase. All features merged into the main Linus tree (even the ones merged in the first week) must be "production" ready (except for new drivers, filesystems and subsystems that are an addition and can't cause a regression). FreeBSD, on the other hand, has a development phase where experimental stuff is allowed and production quality is not a priority - which means that .0 releases take too many to get stabilized and are more buggy compared with a Linux release, because FreeBSD tries to put too many unstable stuff at the same time. It's how Linux used to work in the past, and they had to stop doing it because Linux was way more contributors than freebsd and the whole thing became unmanageable with that model

Re:Well, it's open source, so fork it.

[Anonymous Coward]

btrfs is not even in the same league as ZFS. ZFS is a LVM and fs replacement, done so data doesn't get lost between those two layers.

btrfs offers nowhere near as many featres. ZFS has 64 bit CRCs (which are EXTREMELY useful for finding changed files on backups.) btrfs has 32 bit CRCs which are almost useless as a way of detecting changes, unless one goes by timestamps alone. btrfs also doesn't have transactions (better hope your UPS is up to snuff), and cannot detect corruption on the fly.

Finally, btrfs has not seen any production use and abuse. No way I'm trusting my data to this filesystem for at least 1-2 years, and by then, there will be a "real" filesystem that is on par with ZFS. At best btrfs is a transitional filesystem, like ext4. It isn't a generation changer like ZFS.

Re:Well, it's open source, so fork it.

[lambent]

[citation needed] is not a substitute for meaningful discussion and rebuttal.

"Linux is definitely faster and more feature-rich than FreeBSD." Keeping in the spirit of your post, would you care to post some benchmarks concerning the speed of linux vs. BSD in data storage, or for ZFS vs. btrfs?

at any rate, isn't stability more important in terms of this type of storage? if you're using a NAS-type device, i can't see how speed would be your primary concern, since you're limited by the NAS-style architecture right out the gate.

Re:Well, it's open source, so fork it.

[hedwards]

That requires a citation. Of course Linux is feature rich, it uses a million 3rd party utilities and installs them whether or not you want them. Also, you're going to cite evidence of the statement that Linux is faster if you're going to demand a citation that FreeBSD is faster. Additionally, the BSD license is something that a lot of people view as an advantage, makes it far less of a pain in the ass for companies to help with than the GPL is.

As for btrfs, just let it die, we already have ZFS, Linux has a large number of filesystems supported, but the vast majority of them are pretty mediocre and adding btrfs is pointless when pretty much everybody else seems to be hopping on the ZFS bandwagon. Sure at the moment Apple has pulled ZFS support from being included, but they'll add it eventually. Adding filesystems just to be GPL is an asinine waste of developer talent. Looking at wikipedia's comparison, I'm not seeing anything that btrfs can do which ZFS can't. Definitely nothing worth fragmenting the interoperability for.

Re:Well, it's open source, so fork it.

[hedwards]

OK, now that's just an outright lie. Every OS has a production period where things are merged in and tested for stability and reliability. Linux code doesn't come out fully formed. On top of that, most of what people think of as Linux is produced independently amongst a large number of projects. I've run current in the past and it was hardly as unstable as you're making it out to be. In fact I've experience periods with Ubuntu where that "stable" release was crashing more frequently than FreeBSD current.

I've run Linux in the past, and it just isn't as good as you say it is. I went through a period where I had to reinstall the entire OS just about every reboot because the filesystem was getting horribly corrupted each time it crashed. I'd have to reboot in the middle of the installation because the Ubuntu installation program couldn't handle partitioning in a sane way without doing so. And at the end of the day, I'd have a hodge podge of programs that made up the userland which may or may not play well with each other next time I updated them for a bug fix.

Yes, Linux isn't an abomination and is perfectly fine for many uses, but it's that sort of insulting crap about the glowing development process that makes me not want to run Linux on any of my computers. It's also a pretty blatant lie that FreeBSD changes more than Linux distros do. Over the decade that I've used FreeBSD, Linux has changed far more, and the changes to FreeBSD have mostly been related to the hardware architecture changes that have gone on, in terms of the userland and things that people actually work with, that's stayed relatively constant over that time.

Re:Well, it's open source, so fork it.

[Anonymous Coward]

http://www.phoronix.com/scan.php?page=article&item=freebsd8_benchmarks&num=7

Re:Well, it's open source, so fork it.

[ultrabot]

As for btrfs, just let it die, we already have ZFS, Linux has a large number of filesystems supported, but the vast majority of them are pretty mediocre and adding btrfs is pointless when pretty much everybody else seems to be hopping on the ZFS bandwagon.

"We" here means a very, very small subset of open source community. Ditto for your "everybody else".

Re:Well, it's open source, so fork it.

[Anonymous Coward]

Btrfs is a btree based filesystem, it may not be ready yet but it is far from a transitional filesystem, it may not be a generation changer but that's just because reiser was years ahead of his time. ZFS may have a mature implementation and a shitload of features but it's design is not better than btrfs/reiser4 (I'm not saying it's worse, both are considerably better than standard inode or table based filesystems)

Re:Well, it's open source, so fork it.

[Hognoxious]

"Not only that, but FreeBSD is a far more reliable and higher-quality core than even Debian could ever hope to be."

[citation needed]

What does netcraft say?

Re:Well, it's open source, so fork it.

[TheRaven64]

The 5.x branch was never given the stable designation. The first stable release after the 4.x series was 6.0.

Re:Well, it's open source, so fork it.

[TheRaven64]

Not sure why this was moderated as insightful. FreeBSD has three active development branches. New stuff goes into -CURRENT. After it's been tested a bit, it goes into -STABLE. This is generally pretty stable and is widely tested. Once it's been tested in -STABLE for a while, it goes into a -RELEASE branch. Code is constantly flowing from -CURRENT to -STABLE and from -STABLE to the latest -RELEASE branch. There's no mad rush to move everything from -CURRENT to -STABLE prior to a release.

Re:Well, it's open source, so fork it.

[evilviper]

Well I'm glad you're here to finally save the day and free all those big businesses relying heavily on Linux on their servers with no problems from their OS that apparently drops data like a quadriplegic juggler.

I re-read his post, and yet still missed the part where he said anything about Linux dropping data.

He simply said FreeBSD is higher quality, which is of course endlessly debatable, but may well be true. While your snarky response dismisses any question of software quality out-of-hand...

The same approach could be used against Linux just as well:

Well I'm glad you're here to finally save the day and free all those big businesses relying heavily on Windows on their servers with no problems from their OS that apparently drops data like a quadriplegic juggler. Thanks, anonymous Linux fanboy, the world would be a worse place without you.

Re:Well, it's open source, so fork it.

[evilviper]

That benchmark is demonstrating only the speed of the respective filesystems... The FreeBSD graphs are actually benchmarks of UFS, not the underlying operating system. The Linux benchmarks are EXT4 filesystem benchmarks (nice to see an unstable file system compared to stable file systems). And the OpenSolaris benchmarks are ZFS tests.

Now, if both FreeBSD and OpenSolaris were benchmarked with ZFS, we might have something useful to look at.

And in any case, a single benchmark doesn't justify the claim that one OS is faster than another. There have been innumerable comparisons of Linux and FreeBSD, both having their chance in the spotlight.

Re:Well, it's open source, so fork it.

[jonadab]

> Not only that, but FreeBSD is a far more reliable and
> higher-quality core than even Debian could ever hope to be.

If that's true, it's only because FreeBSD refuses to include anything in the core. Even extremely basic things like Perl and bash are ports-tree stuff and go in /usr/local. I'm sorry, but that's cheating.

Re:Well, it's open source, so fork it.

[Cyberax]

Read it, there are various benchmarks there. And FreeBSD loses in all but one of them. Sometimes badly.

Re:Well, it's open source, so fork it.

[evilviper]

Linux is definitely faster and more feature-rich than FreeBSD.

FreeBSD surpasses Linux in performance benchmarks all the time.

"Feature-rich" is a bit too vague to argue, but I have yet to find software which works on Linux and not Linux... Even Linux binaries can be run on FreeBSD. So I'm at a loss to guess what "features" Linux may have, which isn't found in another OS that can do pretty much everything Linux can...

Re:Well, it's open source, so fork it.

[evilviper]

Read it, there are various benchmarks there. And FreeBSD loses in all but one of them.

What the hell are you talking about? If you're referring to just page #7, what I've said applies...

If you're referring to the other benchmarks they performed, FreeBSD significantly surpasses Linux as often as not.

Re:Well, it's open source, so fork it.

[deek]

Not only that, but FreeBSD is a far more reliable and higher-quality core than even Debian could ever hope to be.

Experience tells me otherwise. I'm in the position where I administer both Debian and FreeBSD systems. This is roughly over the last five years. The Debian systems haven't skipped a beat during that time. The FreeBSD boxes are generally fine as well, although I did have one crash on me once. Luckily it wasn't a busy system.

Five years of running, and not one single kernel or base OS problem. Seems like Debian is fine to me.

Re:Well, it's open source, so fork it.

[LWATCDR]

I am a Linux user but for a NAS ZFS is a HUGE advantage and btrfs isn't here yet so that doesn't matter right now.
Linux is faster? For NAS all you really need to worry about is IO. To be anywhere close to fair you would want to benchmark two FreeNAS and say OpenFiler.
Feature rich? What features do you need outside of a filesystem and networking?
Stability and security are all that matters for a NAS.
The one benefit I see with going to Linux is that it will be easier to integrate into a Linux shop than BSD is.
I see this as more of a marketing move than anything. FreeNAS will now perfectly integrate into an IT shop that is using Ubuntu server.

Re:Well, it's open source, so fork it.

[Cyberax]

"If you're referring to the other benchmarks they performed, FreeBSD significantly surpasses Linux as often as not."

Yes, I'm referring to other pages. You're reading benchmark results incorrectly. Sometimes lower benchmark result is better and sometimes lower value is worse.

Re:Well, it's open source, so fork it.

[mysidia]

The move to Linux IS a fork. It's explained in this artcle about Rumors of FreeNAS' death greatly exagerated

Today Olivier Cochard-Labbé has made a great announcement, FreeNAS will live on and production ready ZFS support will be added with the upgrade to FreeBSD 8.0. At the same time a new Linux version of FreeNAS will be created called OpenMediaVault!

...
Olivier explained it like this: FreeNAS needs some big modification to remove its present limitations (with one of the biggest being the lack of support for add-ons/plugins). We think that a full-rewrite of the FreeNAS base is needed. Therefore, we will take 2 different paths:

1. Volker will create a new project called “‘OpenMediaVault” based on a GNU/Linux and use all his experience acquired with all those nights and week-ends spent improving FreeNAS during the last 2 years. He will still continue to work on FreeNAS

2. And, a great surprise: iXsystems (http://www.ixsystems.com/), a company specialising in professional FreeBSD systems has offered to take FreeNAS under its wing as an open source community driven project. This means that they will use their professionals FreeBSD developers to better FreeNAS! Their manpower will permit a full-rewriting of FreeNAS.

Olivier also added that he will personally come back to actively working on FreeNAS and begin to upgrade it to FreeBSD 8.0 (which is “production ready” for ZFS).

Re:Well, it's open source, so fork it.

[Cyner]

I'm a fan of FreeBSD too, but don't let your zelotry cloud your judgement. The FreeBSD development process is geared toward producing "correct" code as opposed to "feature rich" code. This is apples and oranges, you can't say one is "better" than the other. What you can claim is that one is more suitable to a well definied set of requirement than the other.

Re:Well, it's open source, so fork it.

[mysidia]

It doesn't matter how long you wait between releases, if the quality of your userland is poor.

Junk in, junk out.

Holding releases or releasing slowly does not improve code quality or eliminate bugs. Although it can give you more time to find certain issues before releasing, certain issues will only be found by your users --- and then, your slow release process hurts you (means it takes longer to fix bugs, since releases are so far apart).

Re:Well, it's open source, so fork it.

[mysidia]

Well, Linux does have problems dropping or corrupting bits, and can't detect corrupted bits like ZFS can, which means that silent data corruption is possible.

It's not necessary that Linux causes the corruption. The problem is corruption occurs that Linux can't fix / deal with reasonably, or the way Linux filesystems work is high-risk, and ext3 fsck is horrible, in that data loss or inconsistency of I/Os results and cannot be corrected, despite the fact the filesystem is "journaled": in fact, ext3 journaling is not true journaling, by default, ext3 operates in the less-reliable journal_data_ordered, and not journal_ordered; thus, only metadata is journaled, and data corruption is likely if power or access to storage is disrupted.

The ext3 / ext4 filesystems do not live up to the "journalling" promise.

It's happened too many times to count, that I have lost important data, database, or even entire systems due to Ext3 shenanigans, on Debian Etch, Ubuntu, Redhat Enterprise Linux 4, and Redhat Enterprise Linux 5.

Usually how it happens is a kernel panic, power outage, reset (due to system lock-up), or something of that nature occured, and upon boot, the journal is aborted -- however, there is still data corruption after abort of the journal, or even, the abort of the journal fails, and it becomes necessary to manually run fsck, which "tinkers around with the filesystem" trying to make the metadata consistent again.

I have yet to ever have any issues with ZFS-based or NTFS-based systems; they handle it seamlessly. Windows 2003 or 2008 may show a blue screen once every 2 or 3 years, but the system doesn't require an expert to travel out and manually run chkdsk (Windows equivalent of fsck). And ZFS handles this quite elegantly....

By the way, even NTFS is ahead of ext3 in some ways, in this regard, as far as self-healing is concerned.

Just because it runs Linux doesn't mean the physical disks are magically immune to common issues that effect all storage.

Linux isn't the worst choice for a NAS. (In fact, I would be really scared of the idea of using NTFS for a NAS).

However, it's just as dishonest to suggest Linux Debian is rock-solid for a NAS as it is to suggest Windows 2000 is.

FreeBSD or Solaris is a really good choice for a NAS.

Systems that utilize certain versions of the Linux kernel can be a good choice, with the right configuration, and right supporting programs installed.

For example, a dedicated NAS appliance, can easily provision its ext3 filesystems using SAFE journalling options, instead of the defaults.

And can also pick just the right kernel version and library versions to be as stable as possible under the supported hardware list...

It's not necessarily good for a NAS to support as much hardware as possible: The Number 1 cause of system crashes is faulty hardware or faulty hardware drivers.

It's best to stick with old hardware and driver code that has been around for 5+ years, without major bugs detected.

And to have kernels with isolation features such as IOMMU and NX bit.

Re:Well, it's open source, so fork it.

[Rob Riggs]

Netcraft confirms it: FreeBSD is dying. (Again!)

http://toolbar.netcraft.com/site_report?url=http://www.netcraft.com

Doh!!

Re:Well, it's open source, so fork it.

[JonJ]

He's not calling it stable, he's calling it released. It was posted and isos were available clearly for installation on the FreeBSD web pages, if it was not meant for installation by 'mortals', maybe it should've stayed in the cvs tree.

Re:Well, it's open source, so fork it.

[TheRaven64]

If you went to the front page of the FreeBSD site during that time, you saw two releases. One was advertised as the production release, one as the new technology release. If you were deploying on a production system, you should use the 4.x series, if you were interested in testing new technology, you should use 5.x. I ran 5.x on my laptop, but 4.x on servers. I never had any issues with the laptop, but some people who ran 5.x did. The 6.x series is based on an evolution of the 5.x code, and was considered stable. If you go to the site today, you see 8.0 as the production ready release and 7.2 as the legacy release. The 5.x series was never marked as a production ready release.

Re:Well, it's open source, so fork it.

[Sxooter]

And you're so sure of your experiences that you post as AC. Sure sure, you're not trolling. I've had / currently admin linux servers (file, web, app, db) with years of uptime and no crashes and average load factors of 4 to 40 in daily use. NO CRASHES

Sorry you linux-fu and troll-fu suck so much ass.

Re:Well, it's open source, so fork it.

[rycamor]

> Not only that, but FreeBSD is a far more reliable and
> higher-quality core than even Debian could ever hope to be.

If that's true, it's only because FreeBSD refuses to include anything in the core. Even extremely basic things like Perl and bash are ports-tree stuff and go in /usr/local. I'm sorry, but that's cheating.

No, that's sensible system design. Core is /usr, and has a wealth of system-level programs and utilities, but application software goes in /usr/local. Bash is not the FreeBSD default terminal software; tcsh is. Perl is not a core system scripting language, but it is of course needed for many tasks once you start extending the system in different directions. And /usr/local is not (as in Linux) a little tacked-on directory with nothing in it. it is a very thoroughly laid-out directory tree with everything in a sensible place. Once I moved from Linux to FreeBSD, I loved how everything was in a consistent and unsurprising location.

Re:Well, it's open source, so fork it.

[rayvd]

Agree that ZFS is "way ahead" of btrfs (and most any other filesystem out there other than WAFL really) due to the fact that it's available "now". btrfs will most likely address the issues you bring up by the time they are stable, but still will need a year or two of burn-in time. ZFS needed the same...

I have no doubt that btrfs or something else will eventually fill the void in the Linux world. I think Sun/Oracle does themselves a disservice by licensing things such that ZFS cannot be included in the Linux kernel... all they're doing is ensuring that the Linux community will come up with something compareable down the road when instead they could just let everyone use ZFS and become the de facto standard.

Ah well.

Re:Well, it's open source, so fork it.

[msoftsucks]

Wake on lan for one. I've been using FreeNAS for awhile, and it's been pretty good. The only sticking point for me has been the WOL features which has been a big sore spot in FreeBSD. Having a NAS that I can't wake up remotely is a big disadvantage considering who the target audience for this software is.

Re:Well, it's open source, so fork it.

[badkarmadayaccount]

Wait, what's wrong with NTFS? Not Windows, just the fs. I've always thought it's the best current gen fs, usual host platform notwithstanding. Just check the NTFS-3G benchmarks, and you'll see my point. Meanwhile, I await to see yours.

Re:Well, it's open source, so fork it.

[mysidia]

NTFS gets badly fragmented, as in 80% file fragmentation frequently under many types of load, and the fragmentation of NTFS effects performance substantially. And NTFS doesn't provide self-defragmentation. 40% fragmentation is "normal" for a NTFS filesystem.

It can only be fixed by running a manual tool 'defrag', which is a dangerous procedure (although it _has_ become a bit safer over the years), it's still potentially troubling, degrades performance, and requires lots of free disk space to work properly.

Linux ext3 or BSD FFS is considered severely fragmented if you get 15% of fragmentation. It can happen, but it's exceedingly rare, under almost all common load scenarios, fragmentation never becomes a real performance issue with FFS or ext3, even at such high levels. But that's a side-issue, really:

Well, the main problem is with NTFS on Windows, due to Windows' security issues, and bloat. Windows includes complex libraries and software that are unnecessary for a NAS, and frequently present stability or security issues.

For example: the (very bloated) console GUI, DirectX, Internet Explorer, Windows RPC services.

Some services that may be useful for some types of NAS apps however: mainly file sharing with Windows desktop PCs, e.g. LOW-END NAS uses. However, Eg.: I don't think Windows Storage server is a good choice for a NAS that needs to be highly available (for example, to run virtual machines), and needs to serve NFS and iSCSI.

Stable NTFS-3G is fairly young (~2 years), however, and, i'm not sure that it's as stable as the Windows implementation.

Re:Well, it's open source, so fork it.

[evilviper]

You're reading benchmark results incorrectly.

Apparently so is the guy who wrote them...

"Both FreeBSD 7.2 and 8.0 had also lower CPU utilization than the two Linux distributions tested"

"FreeBSD 7.2 came in slightly behind FreeBSD 8.0 while Ubuntu 9.10 came in fourth and Fedora 12 took a distant fifth place finish."

"FreeBSD 7.2 was slightly faster than FreeBSD 8.0, but both were faster than Ubuntu/Fedora and OpenSolaris."

"OpenSolaris 2010.02 did the best followed by the two FreeBSD releases. Fedora and Ubuntu were in last for this image-processing task."

"Ubuntu 9.10, then the FreeBSD releases, and then Fedora 12."

Read it, there are various benchmarks there. And FreeBSD loses in all but one of them.

I strongly suggest you read it yourself. I also suggest you look at MORE THAN ONE person's benchmarks... there are plenty out there.

Re:Well, it's open source, so fork it.

[badkarmadayaccount]

Here's someone who didn't google something. It's about as production ready as an fs can be.

Re:Well, it's open source, so fork it.

[mysidia]

And that's why it needs a recent bugfix to stop it from crashing and corrupting filesystems on fragmented files?

That's not an exception to the rule. NTFS-3G doesn't have a comparable track record to Linux ext3 and Windows native NTFS implementations.

Re:Well, it's open source, so fork it.

[badkarmadayaccount]

Oops. My bad. Sorry for being snarky.

Huh?

[ToasterMonkey]

Release 0.6x:
- User authentication I must add at minimum LDAP authentication... For NIS and RADIUS I must check if it's possible (don't know if it's possible to use PAM for samba).

Release 0.7x:
- Migrate to FreeBSD 7.0 (with ZFS support)
- Testing a new way for configuring/using share:
'Adding a new disk' will automatically initialize it (format under UFS) and mount it (transparent process for the user).
. 'Creating a share'(create a folder on a selected disk), with user/group/quota property on this share

Release 0.8x:

- Adding monitoring features (SNMP, email alerting, etc..) - Adding other features (I18n Web GUI, LCD, disk encryption, etc...)

Release 0.9x:

- Only Bug fixes, no more new features - This step will depend a lot's about the development of the "geom vinum tools". If this tools is not stable at this moment, I will replace it by 'geom mirror' for RAID 1 and by 'geom stripe' for RAID 0.

Release 1.0:

- The D day! - Lot's of documentation: User guide and developers guide.

and...

Date: 2009-09-17 17:23
Sender: votdev
--- cut ---
Anyway, 0.7 seems to be the last version of FreeNAS as it is right at the moment. For the next version the whole system will be recoded (what i'm doing at the moment). There will be no more embedded installs anymore, also the OS will be Debian.

Regards
Volker

By any other definition, this would be a fork. It's not even FreeNAS any more, it will be CoreNAS?
Anyone have more insight into what's REALLY going on with this project?

Re:Huh?

[Fez]

This story is not the "whole" story.

Basically the author of FreeNAS is going to start over doing it on Linux, but some other group is taking over the FreeBSD portion of FreeNAS:

http://www.freebsdnews.net/2009/12/05/freenas-ready-step/

Re:Huh?

[Cheeze]

He could be creating a deb that could be installed that installs and configures various services to make a debian-based system look like the old FreeNAS. I've used FreeNAS and there is really no reason to have a dedicated system for services that it provides. Many of the people that already use it either use FreeBSD or Linux on another system. If those systems could also run the easy configuration of FreeNAS, they could consolidate systems in their environment.

That's all guessing though. I really hope it ends up being an apt-get package in a repository some where. I like FreeNAS, but I would rather dedicate hardware resources to something else that is more utilitarian, even if that means I have to configure every service myself.

Re:Huh?

[Anonymous Coward]

I propose forking to create a POSIX version. We can call it PNAS. Hopefully that will satisfy at least half the population. I hope nobody forks PNAS, since that would hurt.

New project

[nOw2]

Isn't the real solution to start a new project for a Linux-based NAS solution and leave FreeNAS development to those who want to use FreeBSD?

Re:New project

[frieko]

Isn't that a lot of wasted effort compared to a fork?

ugh

[nomadic]

Why downgrade?


Aww I'm just messing with you all. Anyone who had a genuine emotional reaction to the above needs to go outside right now and recommune with nature.

Re:ugh

[grub]

Try ZFS, which isn't available on Linux due to licensing, and you'll see why it's a loss. I read there's a hack to use it with FUSE but I won't entrust all our data to some shoehorning of ZFS into Linux just to say "We can do it too!"

Re:ugh

[incripshin]

Native support for ZFS is a good reason to choose FreeBSD over Linux. You can make even your root partition ZFS. The reason ZFS is not in the Linux kernel is due to licensing, though.

I hate the GPL ... so much. I don't mind recommending people use FreeNAS because of the licensing. From now on, I'll tell people to use FreeBSD (or OpenBSD if they are awesome).

Re:ugh

[Plunky]

Native support for ZFS is a good reason to choose FreeBSD over Linux. You can make even your root partition ZFS. The reason ZFS is not in the Linux kernel is due to licensing, though.

And yet, the Linux kernel supports MS-DOS filesystems does it not? The reason for that is that although the original implementation license was incompatible with the Linux kernel, a reimplementation was possible. Is it not possible for ZFS? I suggest that if the code is open enough to be included in FreeBSD, the data structures must be documented enough to have an alternative version written.

I'm not saying its not a lot of work, just that it is possible if the desire is there..

Re:ugh

[wiredlogic]

I hate the GPL ... so much.

You will of course back up this vehement disgust by refusing to associate with anything connected to the GPL's backward, misguided socialism. I suggest you start by refusing to make use of GCC, its entourage of hippie fueled utilities, and all applications created with such tools. Then you can truly stand atop your mountain of smugness and enjoy the the wonderful free toys you have to play with after you have purged yourself of that demon taint. Then again, after that, you might find yourself a bit more productive in a Windows environment atop your tiny sand pile.

Re:ugh

[EyelessFade]

And if the linux-kernel had another licence ZFS would still be incompatible.
The authors of ZFS chose this licence because it was incompatible with the linux kernel. CDDL
It says

Mozilla was selected partially because it is GPL incompatible. That was part of the design when they released OpenSolaris. [...] the engineers who wrote Solaris [...] had some biases about how it should be released, and you have to respect that

Re:ugh

[incripshin]

??? I do not promote GPL. I promote BSD. There is no reason for me to have to now defend my claim that I dislike the GPL.

Re:ugh

[incripshin]

Sure. I tried out ext4 when the developers announced it was stable and lost all my data. ZFS is still a valid reason to choose FreeBSD over Linux. If I am going to use ZFS, then Linux is the wrong choice (key word: 'I').

Re:ugh

[CharlyFoxtrot]

Linux: developing stuff that's like totally going to be better than everything else. No really, when it'll be done in date.getYear()+2 it'll rock.

Re:ugh

[Cyberax]

Well, ZFS was declared 'stable' in FreeBSD only this year. Considering Debian release schedule, btrfs might be 'stable' by the time the Squeeze+1 is released.

Re:ugh

[imp]

Even this article points out that btrfs isn't ready for production, while ZFS is in production systems today. How does that make brtfs better? Does it have a better license? Does it have more potential? Maybe. But that alone doesn't make it better today.

Re:ugh

[Cyberax]

Yet the project needs to have a future. Given that it's still in the pre-1.0 state, it might have been a wise choice.

It's pretty clear that Linux is the focal point of efforts in the OpenSource universe right now.

Re:ugh

[harmonise]

Try ZFS, which isn't available on Linux due to licensing

ZFS is available for Linux.

Re:ugh

[hedwards]

The declaration was largely meaningless, it was called unstable mainly because of a lack of developer time to ensure that any showstoppers or glitches would be fixed promptly. It was more a matter of developer caution than actual problems with the code. Admittedly there were some problems with parts of the code that did justify the warning, but the warning was going to be staying until the developer felt that the resources were to back claims of stability.

Re:ugh

[hedwards]

It's GPL, rather than the CDDL that Sun licensed ZFS under. Which leads to headaches in terms of mixing and matching it into the kernel source. But really, it's largely a waste of time, since ZFS is available in some form for pretty much all the other major OSes at this point, and BTRFS may never be so widespread.

Re:ugh

[growse]

Oracle called, they disagree.

Re:ugh

[TheRaven64]

Btrfs (a work in progress for now) is better than ZFS

You realise that the linked article doesn't say that, right? It says that ZFS suffers from fragmentation (irrelevant with flash or L2ARC) and at btrfs won't so much. In every feature comparison I've seen, ZFS does almost everything btrfs plans to do, and the few things that btrfs plans to do that ZFS doesn't yet do are also under development for ZFS.

Re:ugh

[Eil]

I tried out ext4 when the developers announced it was stable and lost all my data.

Hmm. You moved all of your data to a brand-new file system, with no backups, and were surprised at the results?

Re:ugh

[Cyberax]

Ok. Let me rephrase it: "The design of btrfs is better than the design of ZFS".

And it's true, actually. Read the linked article. Btrfs right now has the most features of ZFS: O(1) snapshots, built-in RAID support, easy administration, extents, etc. I've used it in a test environment with great results.

ZFS wins in stability, but that's only for now.

Re:ugh

[icebraining]

http://kqinfotech.wordpress.com/2009/10/23/hello-world/

... and that sucks

[grub]

We've used FreeNAS at home to feed our PopcornHour NMTs. The ZFS implementation works very well for what we use it for (6x750GB and 6x1.5TB drives in RAIDZ)

Now that 0.7 appears to be the last version based on FreeBSD that means ZFS will disappear with the migration to Linux. No, ZFS on FUSE is not an option; too many layers of abstraction for my liking.

Guess the next upgrade will be to native FreeBSD or OpenSolaris. ZFS is so damn great I'm using the filesystem to decide my next server OS.

Re:... and that sucks

[jlittle]

Well, what's stopping people using NexentaStor (non-free) or NexentaOS (free as in beer/speech)? Better yet, Nexenta is OpenSolaris w/ ZFS, etc, but is an Ubuntu LTS 8.04-based distribution. Its always been the best of both worlds. If you have something using ZFS today, you can export the pool, install Nexenta, and reimport, being back up in minutes.

Re:... and that sucks

[grub]

Oh that's a great point, I've forgotten about Nexenta.

Re:... and that sucks

[quantum bit]

If you have something using ZFS today, you can export the pool, install Nexenta, and reimport, being back up in minutes.

Maybe. (Open)Solaris is a bit pickier about wanting ZFS vdevs to be inside of GPT partitions. FreeBSD is layered on top of their GEOM subsystem, so it lets you put a ZFS vdev on just about anything. If it's inside of a bsdlabel partition, Solaris may not be able to find it to import the pool.

It may very well work. Just be sure to have a good backup just in case :)

Re:... and that sucks

[Pinky's Brain]

ZFS has some nice features, but I don't see any of them as being relevant for a media server. For instance, Linux can create a single logical volume from multiple raids just fine with LVM. Or aufs2 for that matter (Linux only). Which has some advantages over either ZFS or LVM (you can lose one RAID and still have all the data on the other). I don't think it's policy based method of creating an union of multiple writable drives is available anywhere else.

Re:why no ZFS?

[wizardforce]

Yes there is but it looks like licensing is a big issue. The talk surrounding Freenas' transition to Linux and the possibility of using Fuse essentially came to a conclusion that there are issues with using Fuse on Freenas for some reason. It may very well be that the implementation is too unstable at this point in time.

Re:why no ZFS?

[TerminaMorte]

zfs-fuse has horrible performance; I compared Ubuntu 9.10 with zfs-fuse (0.5) and OpenSolaris... Ubuntu could hardly do 15MB/s read/write, while OpenSolaris could easily do 70MB/s.

No, not quite the end.

[lonerman]

The new project called OpenMediaVault is linux based, as iXsystems takes up FreeNAS...

http://sourceforge.net/apps/phpbb/freenas/viewtopic.php?f=5&t=4959

openfiler

[headhot]

i feel like the only think freenas had over openfiler was ZFS. i've been running openfiler for 2 years now and it has been rock solid.

without zfs why not go for the more mature linux based NAS?

Re:openfiler

[norton_I]

Openfiler's web gui is buggy as hell, its local LDAP server option is poorly documented and provides terrible diagnostic messages when improperly configured, and it has no official support for installing/booting from flash. Never trust a product that wants to charge money for the admin guide.

I only tried FreeNAS briefly, and did end up using openfiler, but I would love to see anything beat openfiler.

Re:openfiler

[BikeHelmet]

without zfs why not go for the more mature linux based NAS?

Would that be FreeNAS? :P

Because OpenFiler doesn't run on my NAS. It runs fine on my gaming computer, but it won't run on an old Athlon XP...

Glad it's stable for you, but FreeNAS is definitely more mature.

no it stays FreeBSD

[Anonymous Coward]

http://sourceforge.net/apps/phpbb/freenas/viewtopic.php?f=5&t=4959

"FreeNAS needs some big modification for removing its present limitation (one of the biggest is the non support of easly users add-ons).
We think that a full-rewriting of the FreeNAS base is needed. From this idea, we will take 2 differents paths:
- Volker will create a new project called "'OpenMediaVault" based on a GNU/Linux using all its experience acquired with all its nights and week-ends spent to improve FreeNAS during the last 2 years. He still continue to work on FreeNAS (and try to share its time with this 2 projects).
- And, a great surprise: iXsystems, a company specialized in professional FreeBSD offers to take FreeNAS under their wings as an open source community driven project. This mean that they will involve their professionals FreeBSD developers to FreeNAS! Their manpower will permit to do a full-rewriting of FreeNAS.
Personally, I come back to actively work in FreeNAS and begin to upgrade it to FreeBSD 8.0 (that is "production ready" for ZFS)."

FUSE is not the problem Re:why no ZFS?

[leuk_he]

Some googling and i beg to differ:

So that's pretty conclusive, we have a FUSE filesystem (which also claims to not be optimised) which can pretty much match an in-kernel filesystem.

So fuse is not the bottleneck.

I do not see how fuse makes it unstable. Fuse by design keeps unstable code out of the kernel.

Licensing is a problem.

But if people need the feature of ZFS, it could be supported in its current incarnation. Just not as default FS since it current speed is not up to other FS and .Speed and stablity Should be a main issue in a dedicated NAS.

Defending software freedom is a good in the world.

[jbn-o]

When you say "I hate the GPL ... so much." you should have to explain why; something you have not done with the exception that it won't let you link in GPL-incompatibly licensed code such as ZFS. Perhaps your anger should be directed toward those that license incompatibly with the GPL. After all, as the grandparent poster points out, the GNU GPL has done a lot for you as you "promote BSD" systems. Your hatred of the GPL comes off as though you don't understand what the GPL says or why.

bsd,zfs,fork, bad time?

[itzdandy]

ZFS has just come out with built-in on line deduplication. Isnt this what you would call a killer feature for a NAS distro like this? FreeNAS is moving away from a killer feature like this?

In my experience, debian(linux) isnt going to offer significantly better hardware support to justify this switch. No graphics cards or exotic hardware are typically used for a small NAS server and thats where linux has better driver support.

I really like FreeNAS because it is so lightweight, runs from a flash key, does its job without complaining etc. but I dont see much a future for it with this switch. It is essentially a brand new project going up against Openfiler.

Ill have to change my scheme here and export a deduplicated ZFS share via iSCSI and attach it to my windows server to get the AD integrate de-duplicated fileserver.

Re:Defending software freedom is a good in the wor

[incripshin]

It's not that GPL is the victim of incompatibilities. GPL had a hand in causing these. The philosophy of the BSD/MIT licenses is to give freely. The philosophy of GPL/CDDL is take freely, and give to your friends. Generosity vs. selfishness.

I haven't taken the time to read the GPL, but I generally know what it is about. I have read the MIT and BSD licenses. In the same way, I don't care what the ingredients for some processed food product are or why they are there: there are too many.

Re:Defending software freedom is a good in the wor

[incripshin]

As for needing to explain why, you are wrong. I will recommend software based on the license. You cannot respond to me by asking: 'what are you talking about?'

Re:Defending software freedom is a good in the wor

[jimicus]

I haven't taken the time to read the GPL, but I generally know what it is about. I have read the MIT and BSD licenses. In the same way, I don't care what the ingredients for some processed food product are or why they are there: there are too many.

Do so. Even if it's only version 2.

Version 3 IMO is a clarification and adds a number of clauses to deal with new challenges like patents and embedded devices which only run signed code, but the general intent is broadly similar.

kFreeBSD?

[Anonymous Coward]

Sounds like a perfect opportunity for Debian kFreeBSD.

I think you're still too married to the version #

[jonaskoelker]

One also has to consider [tested and maintained, broad userbase, active community, future maintenance]

Those are all very good points.

All of this goes into choosing a sub-1.0 project for something important.

Now, what does the version number have to do with the above considerations? Are you advocating investigating these issues for some version numbers but not others? So if I cobble something together which compiles and I call it "version 1.0", I can sneak my shoddy code past all your careful considerations?

To me, it sounds like the prudent thing would be to investigate the project for future dependability (however you define it, e.g. as above, or more detailed) no matter its current version number.

Re:Defending software freedom is a good in the wor

[Antique Geekmeister]

Then go read the license, at http://www.gnu.org/licenses/gpl.html, before making historically false statements. I've worked with all the licenses you name, and the Apache license, and various closed source licenses. The GPL wins hands down for insisting that open source work remain open source even after local fragmenting, in order to block the very "embrace and extend" that was done to BSD in the 1980's and that was attempted by Microsoft with Kerberos and Java.

If you only "know what it is about" and have never read it, you're in the same position as the USA after Colin Powell lied publicly to us about weapons of mass destruction in Iraq, and you risk similarly wasting massive resources based on statements by a well-meaning but misled leader.

ZFS works great in Linux

[harmonise]

ZFS works quite well in Linux. I've been using it for over a year with no problems.

Re:ZFS works great in Linux

[A12m0v]

Won't it run as a userspace driver? Imagine the performance hit.

Re:ZFS works great in Linux

[abigor]

It's also roughly five times slower, based on what I've read. A non-starter for serious applications.

Re:ZFS works great in Linux

[Just Some Guy]

Yeah, it's pretty horrible. And I don't really get the point. It's not like FreeBSD is particularly hard to install and configure, and configuring Samba to run on it is identical to configuring Samba to run on Linux. I'm hard pressed to think of a reason why you'd want to run ZFS via FUSE on Linux instead of using the real thing on a similar, well-supported Free Unix.

Re:ZFS works great in Linux

[harmonise]

Yes, it runs in userspace. It's not that slow for me. I can get about 25MB/s out of it, and I'm running an older release. A new stable version was released today and it has a number of performance improvements.

Re:ZFS works great in Linux

[harmonise]

It could be slower. I haven't used ZFS outside of Linux for very long and haven't performed any speed tests. But it works for me. I get about 25MB/s with it. I'm using an older version (0.5.0) and a new stable version was released today that is supposed to be much faster. In any case, since FreeNAS is moving to Linux, then if FreeNAS users want to continue using it their choice becomes a matter of ZFS with slower performance or no ZFS at all.

Re:ZFS works great in Linux

[imsabbel]

25MB/s _IS_ "that slow"

Re:ZFS works great in Linux

[Sxooter]

Without any reference to the hardware, it's hard to say if 25M/s is fast or slow. On my laptop running an external USB hard drive that's quite peppy. On my production file server with 16 fast SATA drives it's godawful slow. But without some form of reference, it's just another number.

Re:Defending software freedom is a good in the wor

[laird]

"The GPL wins hands down for insisting that open source work remain open source"

That is exactly why he prefers the MSD and MIT licenses over the GPL licenses - they have different goals. The MIT and BSD licenses give your software away freely to anyone who wants it, which is what you want if you want your software widely used, and the GPL licenses give your software away only to people who agree to also give their software away for free, which is what you want if you want to promote free software. He prefers the former, and you prefer the latter. He stated his preference, and your reply said that he was wrong, which is (IMO) absurd, in that by definition he knows his own preferences better than you do.

Re:who gives a fuck?

[Almahtar]

Yeah man! Fight the power!

Re:BSD is Dead

[PietjeJantje]

Richard, is it you??

Re:Defending software freedom is a good in the wor

[incripshin]

The GPL does more than insist that code remain open. Any source code that is built with it (not by it) must be open. Stallman went as far as to say that code that assumed a GNU interface should be open source and GPL-compatible, or some such nonsense. Therefore we now have the GPL linking exception. At the very least, I at least see the consequences of licenses, and that isn't useless.

Patents

[bill_mcgonigle]

I'm not saying its not a lot of work, just that it is possible if the desire is there..

It's not a work problem - ZFS is elegantly simple, 6000 LoC or so in its basic form.

The problem is it's heavily patented and you have no rights to those patents if you don't derive your code from the CDDL'ed code, which you can't do with the GPL (but FreeBSD, MacOSX, and the FUSE module did).

OpenAFS

[SgtChaireBourne]

Basically the author of FreeNAS is going to start over doing it on Linux, but some other group is taking over the FreeBSD portion of FreeNAS

I was just looking at FreeNAS the other week. It would be absolutely fantastic if FreeNAS started supporting OpenAFS. A lot of sites using NAS are actually distributed around a single city or several cities. A distributed, networked file system would be an advantage for a lot of activities.

Anonymous FTP for download is fine, it's like HTTP. But dropping FTP for upload should be a priority.

Re-doing FreeNAS is exciting news.

A rose by any other name

[sjames]

That's effectively what's happening, it's just that the project name is following the developer rather than the code base. Apparently a fork from 0.7 that will keep FreeBSD is already announced. It will have a new name. The other branch will keep the name and switch to Debian for the next release.

Re:Defending software freedom is a good in the wor

[sjames]

There's a big difference between preferring A and hating B ... so much.

I prefer GPL personally but feel no hatred or even dislike of the BSD or MIT licenses at all.

Re:Defending software freedom is a good in the wor

[Antique Geekmeister]

Preferences are fine. But he made factual claims about the encumbrances of a legal contract he's never actually read. That's not a preference, that's reasoning from hearsay. And it's fixable in 5 minutes by actually reading the GPL. (It's a fascinating document, especially compared to other, "freer" licenses.) The GPL often seems founded in paranoia about how intellectual property will be misused, but that paranoia is historically well founded. (I remember BSD and AT&T UNIX licensing problems: they were awkward.)

Re:Defending software freedom is a good in the wor

[rtfa-troll]

GPL licenses give your software away only to people who agree to also give their software away for free

you are making the parent's point perfectly. If you actually read the GPL carefully, you'd see that anybody can use GPL software for anything even if they don't agree with the GPL or the principles of free software or anything. The GPL is an ideal license for giving software away to everybody.

The restrictions of the GPL come in for people who want to distribute GPL code and only really matter for those who want to distribute derivatives of GPL code mixed with code which under a different license. Even Microsoft can safely use GPL software without ever agreeing to any of the terms of the GPL whatsoever.

Since these things are explicitly and simply spelled out in the text of the license, it's a perfect example of how assuming things about the GPL just because you read them in an internet forum somewhere isn't the best way to go about understanding the GPL. The great grandparent should just simply go and read the license himself.

Re:Defending software freedom is a good in the wor

[Eil]

and the GPL licenses give your software away only to people who agree to also give their software away for free,

Incorrect. The GPL does not say who you can give your software to or what they can do with "their" software. The GPL only says that you cannot impose any practical restrictions on the redistribution of GPL-licensed code.

For example, you can download GCC and do anything you like to it, including nothing at all. But if you want to sell or otherwise redistribute it, you have to make the source code available under the same terms which you received it.

Re:Defending software freedom is a good in the wor

[Eil]

Any source code that is built with it (not by it) must be open.

Any why is that not fair? If someone writes and releases a library under a particular set of terms, that pretty much spells out how they want that library to be used and distributed. If you can't agree to those terms, find a different one. If you don't like the software's terms of use, take issue with the copyright holder, not the license itself.

Stallman went as far as to say that code that assumed a GNU interface should be open source and GPL-compatible, or some such nonsense.

Well, Stallman is a raving lunatic even to people who think the concept of free software is a pretty good idea. His rantings thankfully have no legal bearing.

Re:BSD is Dead

[yup2000]

BSD rises from the dead... news at 10. BSD sighted at Apple and in the wild!

Its not FreeNAS

[nurb432]

If we have a *drastic* fork like this, its only fair that the linux side of the fork should be renamed. With luck the FreeBSD branch wont get absorbed and become commercial, now that its being 'continued' by a for profit company. But if so, then so be it and ill just roll my own.

Re:Defending software freedom is a good in the wor

[incripshin]

And why is that not fair?

Never said it wasn't. /thread

Zfs or nothing

[ggendel]

Zfs has saved my data enough times to warrant it as an essential part of any NAS system I use and value it's contents. I'm seriously looking at the OpenSolaris based NextentaOS and EON as the only remaining choice for an open source NAS solution. I'm also using Zfs on my Mac OS/X machines after I silently lost data on both NTFS, and HFS+ drives and await the community effort to bring zfs on mac up to the latest codebase.

Re:Defending software freedom is a good in the wor

[Rennt]

The MIT and BSD licenses give your software away freely to anyone who wants it, which is what you want if you want your software widely used

Well thats the theory, but GPL software tends to be more widely used then BSD.

...and the GPL licenses give your software away only to people who agree to also give their software away for free, which is what you want if you want to promote free software.

Well thats just wrong. Most GPL contributors don't give a spit about promoting free software, its just a tool to do the job.

Re:Defending software freedom is a good in the wor

[Rennt]

Conjecture! I mean, I don't have figures to back this up either, but I do know there is shirtloads of code in a Linux distro with an intact BSD/Apache/MIT license. I am not aware of wholesale re-licensing just for the sake of it.

But there is no point saying "if BSD code was restricted from GPL licensing it would be the other way around" because; 1) that is not the way things are and; 2) the whole point of BSD is it does not have these restrictions.

Where BSD code HAS been incorporated into GPL it has been done to comply with the BSD license, exactly as intended by the author. If they did not want their code to be used they would have selected a more restrictive license.

I don't get where the resentment comes from, but I would hazard a guess that it is BSD users (or fanboys if you will) rather then BSD developers who insist on these cyclical arguments.

Why oh why not OpenSolaris?

[jotaeleemeese]

Guys, I don't work for Sun (and I think they are in no position to hire people at the moment :-/ ) but you should really drop by to one of their demonstrations to see what they are doing in the storage arena.

Sun has beautiful (technically speaking) NAS devices that allow people to do detailed configuration and troubleshooting using web based point and click GUIs.

Those devices are based on 2 Solaris technologies: ZFS and dtrace.

You can do snapshots, create and share filesystems, find which machines are being pigs with disk usage and for what reason (data or metadata?) all at the click of a mouse.

The important thing is that it is all done with technologies anybody could use, in other words there is nothing stopping the FreeNAS guys, or anybody else, replicating what Sun is doing to offer a non commercial solution when contracted support is not affordable.

I am sure FreeBSD is a great OS (yeah, really) and I appreciate Debian very much, but if there ever was a case of using the wrong tools for the job this would be a sadly good example.

Re:Why oh why not OpenSolaris?

[ggendel]

I don't work for Sun either, but I agree with many of your points. My homebrew NAS server has been running OpenSolaris since build 49 and ZFS has uncovered issues like a flaky sata cable, and an unstable power supply without losing a single bit of data. In some other systems with similar problems, I found silently corrupted data on drives with ntfs, xfs, and hfs+. I would not go to a non-zfs flle system without some kicking and screaming.

I don't know if OpenSolaris, EON, or NextentaOS is missing any of the features of FreeNAS, but I would think about these as a viable replacement. One advantage is that they will always have the bleeding edge zfs enhancements first, like the new deduplication feature.

Re:Why oh why not OpenSolaris?

[RedBear]

One reason: Hardware.

Last time I looked into OpenSolaris I was very disappointed by the very short and mostly unverified hardware compatibility list. The BSDs are much better off in that department and aren't very far behind Linux. In addition, tools like dtrace have already been ported over to BSD (and Mac OS X, which was originally largely based on FreeBSD). The new FreeBSD 8.0 has just taken ZFS support from experimental to official (which in the BSD world I would consider to mean "production ready even in server environments").

Going beyond that, I would personally class OpenSolaris as much further toward the arcane, confusing world of the original Unixen, with Linux being the least confusing for newcomers and the BSDs somewhere in the middle. Honestly, I've had extensive command-line experience with Linux and the BSDs, and I was utterly lost when I tried to get into OpenSolaris. I couldn't grasp the underlying structure of the Solaris-based system at all. It just didn't seem to "make sense" the way most other operating systems have for me. This "friendliness" is probably largely why Linux is more popular than the BSDs and the BSDs are in turn more popular than things like OpenSolaris. If it's easy for new users and new developers to get a handle on how the system works, you will attract more users and more developers.

I'd rate the three similarly in terms of developmental flexibility, which I believe is one of the main reasons they are creating this fork of FreeNAS with Linux. As an ecosystem, Linux development seems to move more rapidly and be more open to new code than anything else out there. Again, with the differing structure of the BSD development ecosystem, the BSDs seem to move at a much more deliberate pace and new code has to go through a fairly stringent vetting process before it is allowed into the core of the system. Let us not forget that while Linux is merely a kernel, the BSDs are largely complete operating systems being developed under the supervision of a central authority, which I think helps add to the stability of the overall operating system. OpenSolaris as usual would fall on the far end of this scale (in my opinion), where it is still somewhat controlled by the people who allowed it to be branched from the commercial version of Solaris, and has numerous restrictions that reduce the speed and openness with which it is developed.

Again, this is just my subjective personal opinion based on years of experience and observation, but I would use this experience to respectfully disagree with your claim that FreeBSD (or Linux) is "[sadly] the wrong tool for the job". Based on FreeBSD's relatively flexible development cycle combined with its observed code quality over the last couple of decades, I'd venture to say that it is the perfect middle ground on which to base a focused-purpose network file server operating system. Such a system needs to be able to stay compatible with a large set of current storage, networking and video hardware and yet be created from highly stable, quality code that changes slowly enough that there is time for a bit of polish before each release.

Re:Why oh why not OpenSolaris?

[Flagg0204]

While I know your speaking of Sun Storage 7xxxx series, and I do agree the UI is awesome (dtrace + zfs is great) where their product line falls apart is redundancy. I can purcahse a NetApp 20xx series filer with redundant heads for cheaper than i can purchase a sun storage 7xxx appliance which is HA. The replication is still based on drbd equivilant, and only recently has ZFS introduced dedup. Sun has a long way to go before they can hold a candle to the mid-lower end Filers

Re: version numbering

[M. Baranczak]

Java is by far the worst offender here. The current version is simultaneously known as 6 and 1.6; which is still better than the previous one, which was either "1.5", "5", or "Java 2 Standard Edition, version 5.0" depending on who you asked.

Re: version numbering

[Antique Geekmeister]

The current release is actually "1.6.0_17" if you look inside the software bundles. Sun has done this before, when they nominally changed 'SunOS' to 'Solaris' but kept parallel but similarly cross-numbered software releases. But no, I'd say that "Windows 2000" and "Windows ME" win the prize, here, since Windows ME was a DOS kernel and Windows 2000 was the NT kernel.

Re:Defending software freedom is a good in the wor

[r7]

most GPL code, including most of Linux, was simply taken from BSD sources and relicensed GPL. What you would do better to look at are clean GPL implementations

Conjecture

Er no, experience rather, from having worked with both GPL and BSD code for many years.

I am not aware of wholesale re-licensing just for the sake of it.

You're not aware of the Atheros relicensing then?

    http://kerneltrap.org/OpenBSD/Atheros_Driver_Developments

Where BSD code HAS been incorporated into GPL it has been done to comply with the BSD license, exactly as intended by the author

That's a novel way of looking at it, but it doesn't really hold water. The Atheros developers certainly did not want their work published under the more restrictive GPL. They, like myself, want all developers to benefit from our code, not only those who agree to publish deltas under the GPL exclusively. But then the GPL never has been about giving.

    http://lwn.net/Articles/247872/

I don't get where the resentment comes from

Conjecture! ;-) The resentment is in your assumption, it was neither intended in nor implied by the original message. So go ahead and downmod (again) anything critical of the GPL, just be aware that when you do you move Slashdot further from the facts and closer to a popularity contest.

I don't see what the problem is....

[VulpesFoxnik]

http://www.debian.org/ports/kfreebsd-gnu/