Slashdot Mirror
Interview With Reiser4 Author Hans Reiser
An anonymous reader writes "KernelTrap has an interesting interview with Hans Reiser, the author of two revolutionary Linux filesystems, Reiser3 and Reiser4. Reiser3 was the first journaling Linux filesystem. Reiser4 is a complete rewrite that is claimed to offer amazing performance and a new plugin architecture offering semantic enhancements to rival Microsoft's WinFS and Apple's Spotlight. Comparing Reiser4 to WinFS, Reiser says in the interview, "Reiser4 is a much more mature design, representing a 10 year effort"."
"Comparing Reiser4 to WinFS, Reiser says in the interview, "Reiser4 is a much more mature design . . ."."
What is better than having the file system's author make unbiased competitive claims about their own products?
LAST POST!!
A B A C A B B
Berkeley was a lot better than junior high school, but it still involved homework, which deep down in my heart I could never believe in.
I hear you. I always avoided homework as much as possible too.
Bradley Holt
I was wondering over the weekend, on a whim, whether it would make sense to create a cross-platform library that abstracts meta-data/search functionality. Like, it would provide one uniform set of utility functions, and this would turn into calls to WinFS on windows, calls to Spotlight on OS X, and calls to ReiserFS on Linux.
;)
But I don't know enough about WinFS OR Spotlight Or ReiserFS to know if this would be even remotely useful or is just nonsense
Irritable, left-wing and possibly humorous bumper stickers and t-shirts
Does that actually exist?
I thought It was dead...
Let's try to keep our comparisons to real entities...
" Reiser3 and Reiser4. Reiser3 was the first journaling Linux filesystem."
The first? XFS, and JFS.
I am Hans, and this is Franz, and we want to [clap] journal your filesystem.
Ya. Ya. All you little girly men with your FAT and NTFS!
Really, Ya. Makes me sad to see such pathetic file systems!
How does the performance of Reiser4 compare to that of Reiser3, XFS, JFS, EXT2/3, UFS, UFS2, etc., in quantitative terms, for various applications?
Cyric Zndovzny at your service.
When asked about GPL the guy sounds like he thinks it's something unpleasant that we have to live for the time being! Jesus. And I had always had lots of respect for Hans. :-(
Journalled filesystems are so 90s. Everyone is raving about how the new whizbang filesystems of the 21st century are going let you do metadata searches, and harken back to the beloved (?) BeOS. Well, what I want to know is: How do I get to this metadata? Some extra tool? Some right-click option that I have to select every time I create a file? Will all File dialog boxes have to be rewritten, and will I have to manually input all this info?
I'm happier with Google desktop, which can, effectively, search many types of my files, and has a relatively familiar interface (for all but RMS).
Once there's an application which can find all pictures of my dog, or songs with piano in them, and store THAT in the metadata, which I can search somehow, call me. Otherwise I'll stick to ext3fs and NTFS.
Wake me when the revolution starts.
I want to delete my account but Slashdot doesn't allow it.
Here's a link to the page that hides the asshats making the pages super-wide with lame comments.
Does anybody know the correct way to pronounce reiser?
.. welcome our Reiser4 dancing tree's, plug-in based architecture overlords.
-You can't lose karma you don't have.
I've been waiting until it's deemed "safe" to use, but it seems it's going on 2 years now or "not ready yet". I know it's ready when it's ready, but is there a timetable for it? I don't have a fast enough spare box to test it out, and I want to dig the faster FS perf on an SATA harddrive. Keep going Hans!
bad_outlook
--
Is this vague enough for you?
OS/2 actually had file meta deta way back in the days of OS/2 2.0. They called them "Extended Attributes". Unfortunately HPFS was not journaled or transacted, so in the event of system failures, your meta data would get screwed up and you would have to run some goofy utility to fix things.
This is my sig.
Frankly, I am perfectly satisfied with the speed of Reiser3. What I want to see is increased reliability and recovery. Performing fsck --rebuild-tree is a Royal pain and takes an eternity on large partitions and the need to perform this operation is far to frequent.
As for new features, I don't need WinFS like features, rlocate is good enough for searching for now. However, the ability to undelete a file would be a KILLER feature as would snapshotting, think Novell NSS with its undelete and copy on write features.
I recently switched a laptop from Linux with ReiserFS3 filesystems to FreeBSD 5.4 using UFS2 filesystems. The size of the filesystems were the same, and the usage pattern (program development, web browsing, etc.) the same.
The UFS2 filesystems had the feel of being quicker than the ReiserFS3 filesystems. That said, I do not have any numerical data to back this up. However, untarring a large tarball consisting of many smallish files under FreeBSD felt quicker than doing the same under Linux.
Would this difference be caused by the filesystems themselves, or would it most likely be a difference between the Linux and FreeBSD IO subsystems? Would ReiserFS4 be more comparable, if not better than, FreeBSD's UFS2 for workstation-style workloads?
Cyric Zndovzny at your service.
Comparing ReiserFS and WinFS is a bit like comparing Qt and Explorer - nonsensical. They're different things, operating at different levels, to serve different purposes.
Come on, how are the parties involved supposed to carry any credibility when making such a *basic* and *fundamental* misunderstanding - /WinFS is not a filesystem/. They also seem to misunderstand what Spotlight is - again comparing it as a filesystem, when it isn't.
Last time I used Reiser I had to reformat back to ext. The starving problem basically made the kernel freeze when flushing buffers during large streaming writes. Is the Large writes starve reads issue gone yet? When I say large I am referring to streaming 12 gig (hour of DV) in a continuous write.
James
I've used it on / for over a year. All other execelent features aside, I really want attributes as files. It's a powerful feature; one could implement something like del.icio.us in a few lines if each link has it's own tags. transparent compression? You know you want it.
reiser1 and reiser2
did they exist? if so what were they like?
note: i'm known as plugwash most places but i screwd up registering that here somehow in the past and now can't register
Eventually, Reiser4 will allow storing metadata in plain old files, so that no special tools will have to be used. It will also allow queries done straight through the filesystem. In the interview, Hans says that that functionality is about 3-5 years away when fully implemented, but it will be implemented gradually, so some functionality will be available earlier.
To get something done, a committee should consist of no more than three persons, two of them absent.
Hah, I thought not puny man.
That comment is remarkably infantile. "If I have seen further than others, it is because I have stood on the shoulders of giants--DUMB GIANTS THAT SUCK!"
blarg.
Jeremy Andrews: Reiser3 is in the 2.4 and 2.6 kernels. Reiser4 is in Andrew Morton's -mm kernel, aiming for eventual inclusion into the 2.6 kernel. What happened to Reiser1 and Reiser2?
Hans Reiser: Just before journaling got added, one of the programmers put two versions up on our website, and bumped the major version number when he should have bumped the minor version number. I was not willing to go backwards in version numbers to fix it because one should never go backwards in version numbers. Oh well. In retrospect, probably I should have gone backwards. Not doing it now.....;-)
The thread you link to nicely illustrates the political manoevering necessary to get a filesystem accepted into the kernel. This is one good reason why filesystems should be implementable in userspace.
There are so many wonderful things that can be done with filesystems once they can be added from userspace. How about transparently accessing files through SSH or FTP, from any application?
There are various tricks that allow filesystems to be implemented in userspace, such as LUFS and FUSE. Other filesystems (especially the ones that are portable to other systems) pretend to be NFS.
All of these suffer a performance penalty, but I wonder how much that really matters when you're interacting with disks or networks, which are very slow compared to the CPU and RAM anyway.
Many things besides filesystems would benefit from being implementable in userspace, but filesystems are what I personally have thought about most.
Please correct me if I got my facts wrong.
I really would like a metadata driven system. Instead of the traditional file dialog for saving or opening files it would be cool to just specify some metadata and have it thrown on a heap of files. I think this is kind of what winFS is trying to accomplish, but above the filesystem level. Hopefully that is in the future of every OS. And if not, or is some better idea comes along, then I guess some time in the future I will pick up a database implementation book and a file systems book, study up and work on it myself.
"Reiser4 ... representing a 10 year effort"
obligatory comment:
by the time longhorn (vista?) is released, it too will be a 10+ year effort.
Normally you have to release something before it can mature. OTherwise its called development...
/how/ its structured, how devels will be able to use it, how we'll be architecting solutions with Reiser4 plugins, it'd be much appreciated.
Still waiting on that plugin system, thanks. Should be good though. No hurry, but if you could even begin to release some info on
-Lord "I hope I havent missed anything in all these years waiting" Myren
But in either case, I'm not quite certain that its valid to assume that all /. readers agree with everything that either Linus or Hans has to say. There are technical arguments for each point of view. You'll find /. adherents on both sides of the issue.
Aside from which, please don't claim that Microsoft has released a true micro-kernel architecture. They haven't.
People who think they need to implement spotlight, HFS+ xattrs or Reiser should read Practical FileSystem Design (pdf) then just go away and use BeFS instead. :-)]
[Actually, the person who implemented HFS+ xattrs and worked in the Spotlight team was the guy who wrote Practical FileSystem Design, so I think that counts
run Linux... oops, I mean, will 2.6.14 ship with it?
There are 2 other significant players in the FS field that Hans doesn't mention:
XFS (from SGI) and GPFS from IBM.
GPFS has a different focus, but XFS seems to be aimed at solving similar problems as ReiserFS (scalability, high performance, journaling).
Simpy
It was not Hans who wrote that, it was Linus in response to Hans' claims that Reiser4 is important because it can compete with WinFS and Spotlight.
I think this was due to a misunderstanding betweek Linus and Hans. Hans wants VFS changes to make Reiserfs better-integrated in order to make it possible to create things _such_as_WinFS_, while Linus says WinFS will probably be implemented in usermode so it doesn't matter.
Indeed, WinFS is basically a queryable object store built on NTFS, which can store metadata about files on the filesystem (as well as objects that are not files). It will provide features that ReiserFS4 will also provide, but is not really a filesystem itself. NTFS already provides things, either shipping or in beta, that Reiser needs to do, such as file compression, transactions, and metadata. Reiser either does these already, or will be doing them soon, but there's no good interface for them.
The problem Hans has isn't copying WinFS, so he shouldn't be talking about it. The problem is that the old Unix filesystem API doesn't support doing things like starting transactions or querying metadata. And god forbid you should want a standard way of using ACLs that would work across different filesystems!
dom
would welcome our new reiser 4 overlords once it gets integrated into 2.6.x mainline.
The _best_ 3D pr0n -> http://www.hookup3d.com
As far as ReiserFS is concerned, the two issues that seem to pop up all the time are the following:
1) It is supposed to be much faster than other filesystems, but under certain circumstances - when one has lots of small files, I believe.
2) Its usage is fraught with danger - there is a steady stream of horror stories out there about disasters caused by this filesystem.
This is why I haven't touched it yet. What I find worrying is that it's been around for a few years now, and such horror stories keep coming.
NTFS still runs in the kernel. Microsoft's "pure" microkernel architecture was heavily diluted in favor of improved real-world performance after NT 3.51. That's why we've seen so many BSOD jokes over the years.
just trying to date the reference...
A goal is a dream with a deadline
but it is horrible for a large networked filestore. The heirarchies that the secretaries at work come up with are convoluted at best, and it takes a long weekend to even attempt to comprehend the logic of their naming convention. When they lose a file, or forget what it was named, when they last worked on it, but can tell you that it was an ISO file (which in itself is ironic), coupled with the fact that they often change the file extensions on their files to random numbers, or try to change it to .pdf to save it as a pdf file, metadata makes a lot of sense.
You also seem to be mixing up the issue of microkernels and monolithic kernels. Apple's OS X uses a microkernel, but the operating system is still monolitic, so all the important stuff is still part of a single big process; the microkernel is basically just an abstraction layer. Hurd is a microkernel but uses multiple servers, so all the subsystems are separate.
And /. is not one monolitic entity either.
To get something done, a committee should consist of no more than three persons, two of them absent.
The point here isn't that Windows and MacOS are moving filesystems out of the kernel (they aren't) or that microkernel = bad. It's that WinFS and Spotlight aren't filesystems, they're just userspace tools that simulate a database-style filesystem (meaning they still use plain old NTFS or HFS+ on disk). Therefore, the existence of WinFS and Spotlight are not convincing arguments for adopting reiser4, since Linux developers could build a similar tool that works with ext3, reiser3, xfs, etc.
In the future, please try reading and comprehending the article before you leap into slashdot bashing...
0 1 - just my two bits
xfs is/was the file system for >= Irix 6.2 (and maybe an earlier version or two). In other words, it was pretty mature by the time it got ported to linux.
I never said the above words attributed to "Rieser".
I am sure of it becuase I would absolutely never say that "Linux kernel developers do what's right because it is _right_, not because somebody else does it.".
I am just not that nice a guy that I would say such a thing.:-/ I am guilty of saying the opposite at various times. I am known for this, and not particularly liked for it.:-)
This is a forged quote. Note the false defensiveness put into it in the sentence "So there's really no point in trying to push your agenda by trying to scare people with MS activities." That really sounds like someone at MS posted this.
It does not matter so deeply that MS put it into or out of the kernel, what matters is how they layered the code relative to itself --- that is, do they use the FS API, which lacks an insert or excise operation, to repack small objects that they squished together within a file, and does that layering make things slow. I think it almost certainly does make it slow, and it definitely is inelegant.
Hans Reiser
Reiser4 Architect
Namesys
A bit extreme, but I see your point. I've found ext3 and Reiser3 to be of roughly the same stability, but I've had much better success with ext3's recovery tools than Reiser's.
When they're working, they both work great, and they both handle unexpected shutdowns well. But when they crash and burn, you're more likely to be able to salvage a broken ext3 system than a broken Reiser3 system.
I see this comment in danger of getting modded into oblivion, so let me back it up.
I haven't used Reiser for years for exactly the same reason. The two times that I had had any trouble with a reiser partition that required the use of recovery tools, the reiserfsprogs segfaulted, and then destroyed all of my data in the name of "recovering" it. Even e2fsprogs is far more stable (and useful) than that.
His point is that WinFS is in user mode because it's not a filesystem, it's an application that runs in the background.
Yes, that sucks.
The global economy is a great thing until you feel it locally.
Over the past several years, we had pretty good luck with using Reiser on root and data filesystems. Good luck in the sense that we never encountered something we couldn't recover from. However, we did have more than one instance of filesystem corruption that would crash the kernel (We used it on several of our development servers). The warnings on the 'rebuild tree' utility weren't very reassuring, but it always seemed to work. We also had instances of corrupt files by sticking random bits of data of other files at the end.
I'm migrating our servers slowly over to ext3 as we upgrade them, mainly because it is more mainstream and I prefer my source code sunny side up as opposed to scrambled. I noticed that the same number of files seems to take up less room (10% or so?) on the disk overall with Reiser than with Ext3 (as reported by df).
ayershome.org/users/eric
> But in either case, I'm not quite certain that its valid to assume that all /. readers agree with everything that either Linus or Hans has to say.
/. says there's this uber-mind thing, and everyone thinks alike.
Really? But, *everybody* on
I've been thinking about starting up a file system project (as you do), and was wondering if anyone has thought of using something like the FUSE kernel module with a database (say MySQL or Berkeley DB) to create an easily indexible file system. The idea is to create a basic proof of concept using FUSE and if it gets any interest turn it into a proper (kernelspace) FS.
What sort of problems can I expect to face?
Ah that's hilarious, ok, yeah, it was Linus saying it. Apologies for not remembering and thinking it was someone at MS.:-))) Can't imagine why the poster put my name on it though.....
Hans Reiser
Those are the words of Linus Torvalds in response to someone suggesting that Reiser4 should be merged, in order to stay competetive with WinFS and Spotlight. To counter the reasoning, Linus Torvalds stated the following seperate points:
1. WinFS is not the real filesystem, the real filesystem (NTFS) still runs in kernel mode. WinFS is "merely" a set of libraries in user-space, like gnome-storage. So, you can't derive a need to push such functionality into the Linux kernel.
2. Trying to push some functionality into the kernel with the reason to compete with Microsofts development won't work, because they do what they think is right, not in order to compete with someone else.
"Between strong and weak, between rich and poor [...], it is freedom which oppresses and the law which sets free"
The topic in channel #gentoo-amd64 on irc.freenode.net has said "Reiser4 is evil" for more than a year. Does anyone know if Reiser4 actually works in a x86_64 environment?
Exactly.. You'd think a filesystem recovery tool would be more stable than that.
Here I was thinking "oh, it's a journaling filesystem, all I have to do is run this nifty tool and it'll correct any problems, after all, this is why I'm using this filesystem, right?" and the next thing I see is a segfault.
Did nobody else read this quote?
"I will probably keep on doing GPL work for now. It is not an easy life, I am $200k or more in debt and drive a 1989 CRX Si"
$200,000 in debt? I hope that's in pesos. That or he's just counting a mortgage as debt or something. Otherwise he has some serious budgeting problems.
If you wanna get rich, you know that payback is a bitch
Not really looked at GPFS, but if IBM's history is anything to go by (JFS, M:N threading, the DAISY code translator, etc) it'll be revolutionary, be an inspiration to a thousand projects, and get forgotten as it is overtaken.
Sad, but true - IBM has done masses for Linux, in terms of proof-of-concpet, forcing the pace and introducing new idead. Unfortunately, they then drop the ball. It hasn't mattered much, as others have gone running with the ideas, but it would be nice if IBM saw a real return on their investment by keeping on until their technology is adopted.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
This was discussed in Slashdot some time ago. The general consensus seemed to be that they were legit, with provisos. Those being that a more general range of tests failed to show nearly as much gain with Reiser4, but for specific problem-spaces, Reiser4 was undoubtably the best. It just wasn't universally the best. XFS could beat it in some things, I believe, as could JFS.
It's a small world and it smells funny; I'd buy another if it wasn't for the money; Take back what I paid (SoM)
How the hell did that get past the lameness filter?!?
I'm no expert on slashcode, but if that won't trip the filter, what the heck will?
This sig rocks the casbah.
Well, that makes a lot more sense now, and surely the poster was just confused rather than malicious.
Linus and I disagree on this point, a pity that.
last week, I formatted my new USB hard drive with reiser4. It works well with both 32 and 64 bit (kernel and userspace), no problems so far.
Urm, a mortgage is debt.
Mortgage debt is real debt. I assumed he was talking about that. Otherwise, he is seriously screwed.
This is interesting and all...
But why classify this in the hardware section? Wouldn't the Linux section be more appropriate?
perl -e "eval pack(q{H*},join q{},qw{70 72696e74207061636b28717b482a7d2c717b343 637323635363534323533343430617d293b})"
Hans takes it up the ass, much like his faggot zealot buddy (and all slashdot faggot zealots, for that matter) linusux whorevalds does. monkeys could write better code.
Perhaps this is troll/flame bait material but I'll bite...
I've been using ReiserFS exclusively for about four years now and in that time have NEVER had problems attributed to ReiserFS.
1) It is supposed to be much faster than other filesystems, but under certain circumstances - when one has lots of small files, I believe.
Such cases are extreme. IIRC it isn't merely a problem with small files--it is when those small files are all contained within a single directory. They have to be very small files, and many many thousands in number. In practice the ONLY time I've ever encountered such a situation is when running benchmarks--most sane individuals would have the sense to organise such files in subfolders, and one might even re-examine their application to remove the requirement to have such an insane configuration.
2) Its usage is fraught with danger - there is a steady stream of horror stories out there about disasters caused by this filesystem.
As I stated I've NEVER experienced a "horror story" first hand, nor have I even seen any evidence of such except for a handful of anecdotal stories. OTOH, I've seen and personally witnessed MANY distressing situations involving filesystems like EXT2 and FAT (word to the wise--you are REALLY inviting trouble if you are running MySQL--without innodb--on an EXT2 filesystem. If you have a power interruption at just the wrong time you are doomed if you even remotely care about your data).
I do suspect this is flamebait as every statement you make is unsupported. As with any project there are bugs, but "horror stories"? Honestly--point me to ONE single article that is a true ReiserFS disaster--and not one on an ancient, obsolete pre-journalling version of ReiserFS or someone complaining about a test/alpha/beta release eating his data. In my experience Reiser3 as included on all production Linux distros is mature, extremely stable and performs admirably (when I converted a mail server to ReiserFS--all hardware remaining the same--the difference in handling very large mailboxes was amazing).
is that nobody can tell how much time has been spent on it. Also, it draws from previous work. I would find it highly unlikely that a useful mesure of such metrics can be established in this case. Also there is differring of features to be considered. For example, say one browser has tabbed browsing. and rudimentary rendering standards support, and another has support for many standards, but no tabbed browsing, and still another will render standards correctly but also handle content that doesn't conform to standards. Which is more mature?
1. Buy a mac
2. Press command-space
Linus said that, not Reiser.
OK, let's look at actual time invested in each FS over 10 years...
ResierFS = what? 50 man years? 100 man years?
WinFS =~ 10K man years? 100k man years?
Both at 10 years old. But... Some 10 year olds are fairly mature while some are essentially two year olds still running around yelling "No!", just in bigger bodies.
Physical age is no indicator of maturity.
The quote is genuine, but it is from Linus Torvalds. Gee, he really believes that Linux kernel developers do things because those things are right? How'd have thought that?
If you haven't used ReiserFS in years, then you are perhaps not qualified to comment on it anymore, but I'm sure you'll find people who agree with you. Personally, I have recently (july?) had a different experience. I did a fresh install of Debian on a new drive, but after a short while, I got errors on two file systems, perhaps because of bugs in the tools used in the installer, or perhaps because of power outages: / (ext2) and /home (Reiser3). Reiser3 needed a fsck --rebuild-tree, and worked after that with no problems. I tried three different binaries of e2fsck, which all segfaulted; the one on the installer, the one in Debian Sid, and a statically compiled downloaded elsewhere. Eventually, I just copied / to /home, made a fresh / and copied it back, and I've had no problems unrelated to Apples shit implementation of OpenFirmware since then.
Just a story to tell that YMMV.
The trouble with making the filesystem too fancified, is that you end up with 2 namespaces of tool programs - the stuff in /bin and the stuff in filesystem plugins. And, there's not necessarily any sensible dividing line as to what belongs where. He talks about a "childcat" plugin that would merge the contents of a diretory into a single virtual file. But why seperate this from the existing cat command?
/home/reiser/mp3s/..../childcat > /dev/dsp". Two totally different ways of performing two very similar tasks, strung together in a single command line. You'd need to understand both to begin making sense of it - how to call them, how to pass options. You'd need to understand that "cat" was available to the whole system but "childcat" only on disks formatted reiser4. You'd need to learn which tasks were divvied up into what sphere. Can you imagine the difficulties in explaining this all to a Unix beginner?
You can see the confusingness in the command line he gives, "cat
Reiser4 looks nice in many ways, but it does seem to suffer from "if all you've got is a hammer..."
we had problems with amd64 for up until this may when one nice company http://t-platforms.ru/ lended us dual amd64 mobo for couple months which was enough to get rid of couple annoying bitmap operation bugs....
... i guess it works... real evil is an oil war.
so
Regarding Hans's comments about cp, I've recently been wondering why no filesystems seem to support copy-on-write semantics. Instead of physically copying all that data around, just mark it as copy-on-write with a new name somewhere else in the filesystem. Later, if one of the copies is changed, it then automatically gets copied. This could be a nice foundation for certain types of backups and versioning. It's similar to hard-links, but the link breaks automatically when one copy is changed.
To get something done, a committee should consist of no more than three persons, two of them absent.
Building something that is called FS (filesystem) but actually works on top of the real (in-kernel)filesystem and runs in userspace doesn't count as microkernel architecture.
Linux is not Windows
...Hans Reiser, the author of two revolutionary Linux filesystems, Reiser3 and Reiser4.
So who authored Reiser1 and Reiser2? Was it Paul Reiser? Do you have to be named Reiser to work on the filesystem, or was it just a coincidence? Inquiring minds want to know.
If you can read this sig, you're too close.
The uses for semantics, plugins, and massively superior performance are all very compelling. Yes, some of the world needs to be redesigned around it--but some of these redesigns are long past due, anyway.
Breaking backward compatibility--if necessary--is worthwhile and certainly not unprecidented in the kernel (Remember a.out, remember the big conversion from libc to glibc). What it unnerves is some people's warm feelings about Ext3, regardless of technical merit. It's largely, I think, those same people who claim Reiser 3.6 is unstable.... Funny how, on my systems over the years, that's only been the case on Red Hat--not even on Mandrake which is based on Red Hat.
Matthew C. Tedder
We (my company) deliver servers to printing houses in europe, asia and even to the US. Okay: We have had some trouble with reiserfs in the early days (this was with Linux 2.2.14).
I have never experienced serious trouble with systems from 2.4.18 and up that was not caused by a hardware failure. Even than: With reiserfsck --rebuildtree volume it was sometimes possible to recover from some of the damage caused by disk read failures.
Of course I've to admit that 80% of our customers run SuSE Linux and the SuSE kernels always differed somehow from the official http://kernel.org/ kernels.
The performance gain of reiserfs compared to ext3 makes a big difference in our application area. I'm looking forward to find some time evaluating reiser4 and I like the work of Hans Reiser.
Peter Funk, Oldenburger Str.86, D-27777 Ganderkesee, Germany
if you havent used reiserfs in years (as you claim) then your experience is completely outdated and no longer relevant to the discussion.
Urm, a mortgage is debt.
Urm, but usually you can sell your house and be debt-free. And you will be making "debt payments" to your own debt if you own, or the possible debt of your landlord.
WindowsNT is not really a true micro-kernel. And neither is OS X. They both technically use 'Mach' underneath, but they are otherwise largely monolithic OSes. Linux, even though it isn't a microkernel, is actually much less monolithic than either OS X or NT.
And, as Hans pointed out, WinFS is evil not because it's in user space, but because it isn't really a filesystem from the point of view of most programs. It's an add-on API layered over the filesystem, so programs have to do extra special things to use it.
Need a Python, C++, Unix, Linux develop
"It takes more than a license to make code open"
yes, the native hpfs in os/2 did store the EA as part of the fs, not hashed off like it was for FAT.
... i dont this design choice (internal vs external storage of EA metadata) was not due to any inherent limitation in the orginainating platform at apple - because the HFS eventually supported multiple forks, which would have easily supported a 'slot' for something like extended attributes. But that was a while ago, so it's hard to say clearly.
man, those were the days - sweet sweet os/2.
surprisingly, while ibm made many useful improvements in the joint venture with apple on TALIGENT (eg the collection classes in SOMObjects were derived from taligent), one of the issues IBM did _not_ bring to the table at Taligent was it's meta-data experience with os/2.
this was a disappointing ommission.
to counter the perfidious microsoft 'traitors' inside ibm, os/2 was always looking for a "killer app" -- ironic that the building blocks were always right there in the fs staring everyone in the face.
a bit of trivia: surprisingly, the workplaceshell was not orginally mandated to use somobjects; thus the wps might have had to access the meta-data in the extended attributes directly. perhaps becuase SOM did such a wonderful job of abstracting access to the EA, developers never truly appprecaited its elegance & power.
i dont reacall how the BENTO-based compound storage mechanism for OPENDOC made use, at all, of the extended attributes; my recollection of the bento API is that each content model had its own semantics, privately represented inside bento - so perhaps there was no need/ability to publish this information externally in the core fs
Anyone have any more direct recollections?
cheers:dlf
Look, ext3/2 blows, its garbage
I put 2 120gig HDs in my system and ext3 wouldnt mount the 2nd one, so I gave up and use finaly reiser3 and bingo it mounts.
DUDES, if you gona code and make some thing fail, learn how to communicate a decent error message.
That goes to apple too, darn -3136 errors , its not like the HD space is low for 500chars of text.
Liberty freedom are no1, not dicks in suits.
The interview says this:
Is that really anything new? SGI's XFS filesystem, which came out on IRIX 5.3 in late 1994, had special features for storing small files' contents inside the i-node instead of in data blocks. Here's some text from an XFS whitepaper on the subject:
I am not that familiar with the reiserfs history, but did reiserfs come out before 1994? My guess is that it didn't. If not, then perhaps it uses a different approach to storing small files or takes it to a new level, but it doesn't appear to be the first filesystem ever to put special emphasis on storing small files efficiently.
I'm not saying that reiserfs sucks or anything, but it does seem like an evolutionary improvement more than a revolutionary one.
Classic bit of coding wisdom, there.
The first thing to understand is that there is no "/. POV", because there are thousands of different people posting to Slashdot. You are going to see people advocating on both sides of every issue. That doesn't mean there is a double standard, only that there are diverse opinions.
Or, to put it more rudely: if something strikes you as stupid, think a minute before posting and make sure it isn't you.
I don't care if it's 90,000 hectares. That lake was not my doing.
Not some, I pay almost all.:-/ Unfortunately a lot of the debt is to the people whose salaries are overdue. I thank uhoreg for understanding that funding not budgeting is the problem. Budgeting is easy to do, it is funding that is not so easy.
"why no filesystems seem to support copy-on-write semantics"
Look at Sun's "ZFS". It does this. It also supports pointin time recovery. And being transactional you can back it up whioe it is in use. Another great ZFS feature is that it includesthe "storage manager" layer in the FS. So mirrors and raid ad spanning volumes as well as access by a cluster of machines is al in the FS. You can do things lie add a new physical drive to the FS while the system is running. ANd being a 128 bit file system you are not likely to fill it up soon I've read about Sun's testing. Seems they have crashed system intentionally write doing heavy write manymany thousands of times and have yet tosee any data corrution. Now that Solaris is Open Source we may oneday be able to know who this works.
Not working in gentoo hardly means that it's broken.
Besides indexing stuff that has no user-space support yet, I'd like decent recovery tools, at least on par with Ext3, and no memory of Reiser getting into fist fights with such basic stuff as NFS. Does it even work with SELinux yet?
Not really. OS X is about as much a microkernel as NT is - which is to say, in theory but not in practice.
Both are designed around discrete modules providing services via messaging, like a microkernel.
Unfortunately, they also both run just about everything in kernel mode, removing the stability benefit of a microkernel design (in exchange for better performance).
However, I suspect both would be relatively easy to turn back into a real microkernel (move most stuff back into user mode) if the hardware ever gets fast enough (which it might, with multicore CPUs starting to appear) due to their extremely modular designs.
Sounds rockin'. I'll have to look into that. Not that I'll be running Solaris anytime soon, but it's always nice to get some more ideas.
I'm just chiming in with the folks who want to love Reiser (v3) but have had poor experiences. On my big SATA system, I started with three 1.7TB ReiserFS v3 filesystems (3Ware with 8 drives each on RAID5). These get hammered pretty hard for research, for database, and for public anonymous FTP servers.
Eventually, I changed them to ext3 instead. The reason is that when a problem occurred requiring fsck (such as a power outage or kernel panic) it was always the ReiserFS volume that required recovery, and fsck.reiser would take many hours. The ext3 filesystems (once changed over: same hardware, same usage) just didn't have such problems, and when unmounted uncleanly would take just a few seconds to fsck.ext3.
A 1.7TB filesystem (currently kernel 2.6.11 on a dual Xeon box with 12GB of RAM) is still relatively big, so I wanted to share my experience.
I'm ready to try Reiser v4 any time, though. In my benchmarks, it was much faster for dealing with small files, for big directories, and other things that are common on big research systems. But ext3 seems a bit more stable, and as others have mentioned is far faster & easier at recovery from problems.
Can you point me in the direction of your donations page? (if you have one) I've used reiserfs on several computers and its worked out great. So I'm more than happy to help you out a bit when I can.
It shows that they don't know how to build a decent set of tools before "releasing" a filesystem, and that maybe after ten years this one still isn't ready to take out of the oven.
well gee ext2fsck segfaults on me too. i guess that means ext2 isnt ready to take out of the oven either.
hell i can make the _kernel_ panic on ext2 errors, i guess that means linux isn't ready either.
well gee ext2fsck segfaults on me too. i guess that means ext2 isnt ready to take out of the oven either.
;)
All I can say is I've never seen it, and I've put a lot more use into ext2/3 than I ever have Reiser (and yet Reiser's failed more).
hell i can make the _kernel_ panic on ext2 errors, i guess that means linux isn't ready either.
So can I. mount option "errors=panic" or superblock equivalent
Having metadata is one thing. Having a database engine that can actually perform magic with that metadata is another. Unless ReiserFS implements a DB engine as well, then comparisons against WinFS are pointless.
well, we ran ext3 and experienced numerous catastrophic system failures on production servers. our reiserfs servers never experienced catastrophic failures. so we switched. no more failures.
about 30 production servers. enough data points for us. shrug.
While not a donations page per se, you can visit their support page: http://www.namesys.com/support.html where they accept credit card payments for technical support. I'm sure that they wouldn't mind if you dropped some money in there and told them that it's a donation rather than a support request.
To get something done, a committee should consist of no more than three persons, two of them absent.
$200,000? Thats peanuts to a real business. Frankly I'm surprised it is that low. Many small startups burn through more than that per month, and they end up just fine.
Now he has been in business for 10 years, so I would hope that he has a plan to make money. It wouldn't take that many sales to pay off the dept. However that doesn't mean it is easy.
I'd suggest that he needs salesmen, but I'm not sure who they sell to. If they don't have a business plan, they need that. If they have a plan, then it is just a matter of not going bankrupt before the sales kick in. (When and if they will kick in I don't know)
Hans,
Keep up the good work. Despite all the negative comments here, you are doing a fanatasic job. Few remember that until you started work on reiserFS, not one filesystem included in Linux had journaling! IMO, you definately kicked off the Journaling FS movement for Linux.
One important feature, would be to provide an API set that provides journaling at the file level. For instance, a developer might create an app that stores data in a group of files, and if the app crashes or the machine unexpected shuts down, the file data managed by the app might be come corrupted. By providing a ReiserFS4 plug-in and API set for App journalling will be huge advantage.
I am looking forward to working with ReiserFS4 for quite some time. Thanks for all your time and hard effort!
I read through that thread and the only person I saw acting like a contrary child was one Horst von Brand. If you're so emotionally invested in that thread you can't comment without blatant bias then you should leave it to the impartial people to comment.
He was unnecessarily confrontational and appeared to be a class A trouble maker taking any opportunity he could to mis-state facts and misrepresent the parents words in order to ridicule.
This is the independent eye view of a normal man who doesn't have his head firmly wedged up anyones butt displaying dog-like slave-love or have a size 12 chip on my shoulder.
HIV+ semen -- goes down smooth!