RE: [gentoo-dev] reiserfs

RE: [gentoo-dev] reiserfs

[Sean Mitchell]

> I was reading your FAQ on your website talking about reiserfs being
> unstable with 2.4 kernels.  I have run it since 2.4.5 on slackware,
> using it on 2.4.5, 2.4.12, and 2.4.18 now with no problems. What
> problems did you guys experience?

This topic needs to be expanded upon in the FAQ. It seems that once a week
this question is asked, or there's a post telling us all that someone has
been using it for months with no problems so obvioulsy it's all fine.

Sean

Re: [gentoo-dev] reiserfs

[Alexander Gretencord]

On Tuesday 14 May 2002 16:56, Sean Mitchell wrote:
> This topic needs to be expanded upon in the FAQ. It seems that once a week
> this question is asked, or there's a post telling us all that someone has
> been using it for months with no problems so obvioulsy it's all fine.

Well, many people have run it without problems. SuSE even ships it with their 
distribution since ages (before it got into the main kernel tree). Without an 
explanation _why_ you think it's not stable enough the statement is worth 
nothing. So ACK, this definately needs an entry.


Alex

-- 
"They that can give up essential liberty to obtain a little temporary safety 
deserve neither liberty nor safety."
Benjamin Franklin

Re: [gentoo-dev] reiserfs

[Jean-Michel Smith]

On Tuesday 14 May 2002 10:07 am, Alexander Gretencord wrote:

> Well, many people have run it without problems. SuSE even ships it with
> their distribution since ages (before it got into the main kernel tree).
> Without an explanation _why_ you think it's not stable enough the statement
> is worth nothing. So ACK, this definately needs an entry.

First, an appeal to authority (Suse in this case) is a logical fallacy you 
should not engage in.  Just because Suse ships reiser with their distribution 
doesn't make it stable or safe for production use.

In fact, a friend of mine who runs a computer consultancy, and uses Suse in 
nearly all of his installations, was bitten very badly by a reiserfs bug that 
resulted in near-catastrophic data loss.  I say near, because he was able to 
recover from backup tapes.  Nevertheless it resulted in an allnighter getting 
the system back up, on a more reliable ext2 filesystem, followed by several 
days work as he moved other installations off of reiser and onto ext2 (his 
choice, not mine ... I probably would have opted for JFS or ext3 in his 
particular case).

I have personally witnessed data loss using reiser on numerous occasions ... 
symptoms ranged from strange "undeletable" files that were corrupt, to entire 
directory trees vanishing for no apparent reason (but the disk usage 
remaining unchanged).  No recovery was possible in either case (short of 
reconstructing a new filesystem from scratch and restoring from backups).

In all these cases all of us had all been using reiserfs "for months with no 
problems" ... and we still suffered severe data loss.

Reiserfs is NOT ready for production use, and the gentoo FAQ is both wise and 
friendly for pointing that out and guiding people away from that particular 
folly.

There are plenty of other, much safer filesystems to use, including XFS (if 
you don't need bleeding edge experimental features, e.g. can be happy with 
stock 2.4.18 kernel + xfs patches), JFS, ext3, ext2 (no journalling), and so 
forth.  I know people tend to get very emotionally attached to whatever 
filesystem they like, but this IMHO is unwise ... one should be very agnostic 
about what fs one chooses to use, rather than defending a particular choice 
"to the death" as seems so common with software these days.

That having been said, there is a plethora of hard evidence as well as 
anectdotal experiences to learn from, and to indicate that reiser really 
isn't a safe choice to be making.  This is reflected in the gentoo 
installation documentation, IMHO exactly as it should be. 

Jean.

Re: [gentoo-dev] reiserfs

[Mark Bainter]

Jean-Michel Smith [jsmith@kcco.com] wrote:
> recover from backup tapes.  Nevertheless it resulted in an allnighter getting 
> the system back up, on a more reliable ext2 filesystem, followed by several 
> days work as he moved other installations off of reiser and onto ext2 (his 

And I've spent all-nighters restoring from ext2 filesystems, simply because 
the system wasn't shut down properly.  

> I have personally witnessed data loss using reiser on numerous occasions ... 
> symptoms ranged from strange "undeletable" files that were corrupt, to entire 
> directory trees vanishing for no apparent reason (but the disk usage 
> remaining unchanged).  No recovery was possible in either case (short of 
> reconstructing a new filesystem from scratch and restoring from backups).

I just had strange undeletable files on my ext2 filesystem less than a 
week ago.  In fact, once I finally got it fixed, it happened two more times
that same night.  

You also suggested ext3.  Ext3 has been a disaster for me thus far.  It's
been the only filesystem I've had trouble with (outside of the typical
issues we're all used to with ext2).  However, I wouldn't go so far as to
say nobody should use it just because I and a few people I know have had
bad experiences with it.  I know there are people out there running it 
successfully but I wouldn't even trust it to hold my /tmp filesystem.

> In all these cases all of us had all been using reiserfs "for months with no 
> problems" ... and we still suffered severe data loss.
> 
> Reiserfs is NOT ready for production use, and the gentoo FAQ is both wise and 
> friendly for pointing that out and guiding people away from that particular 
> folly.

I just can't agree.  What exactly is your required time frame for running 
reiserfs with no problems before you think it's stable?  I personally have
been running reiserfs on my systems since before it was even merged into the
mainline kernel.  I work the hell out of my systems and I've never had a
problem.  

I've had it in production systems as well, for almost as long.  In systems
ranging from large, high load email servers, to web servers handling approx
1.5 million page views a month.  I've never had a problem with it.

I'm not trying to be insulting, but lets not forget the human factor in this
equation.  I don't really know you, or the people you cite, so please don't
take this as a slam on your skills, I'm just noting there's more to factor
in here than just the filesystem.  If you don't know reiserfs that well, and
just know ext2, or ext3 maybe you have less problems because of that, rather
than because of the relative quality of the two filesystems.

Re: [gentoo-dev] reiserfs

[Jean-Michel Smith]

On Tuesday 14 May 2002 10:52 am, Mark Bainter wrote:
> Jean-Michel Smith [jsmith@kcco.com] wrote:
> > Reiserfs is NOT ready for production use, and the gentoo FAQ is both wise
> > and friendly for pointing that out and guiding people away from that
> > particular folly.
>
> I just can't agree.  What exactly is your required time frame for running
> reiserfs with no problems before you think it's stable?  I personally have
> been running reiserfs on my systems since before it was even merged into
> the mainline kernel.  I work the hell out of my systems and I've never had
> a problem.

First, I've had resier lose data on systems that were running fine, i.e. were 
NOT shut down improperly, or suffered a kernel hang, or any other sort of 
disruption that one could reasonably expect would lead to filesystem 
corruption.

In all the years I've been using GNU/Linux (since 1993) I have never seen this 
on ext2.  Nor have I seen it on XFS (which I have been using for over 3 years 
now on several production boxes).  I have not seen it happen with JFS or 
ext3, though admittedly I haven't used either of those two nearly as 
extensively as I have ext2 and XFS.

So, in answer to your first question, I require that a filesystem NOT 
spontaneously lose or corrupt data, or mysteriously delete entire directory 
trees with no apparent cause.  To date all of the filesystems I have tried 
have met this rather modest standard, with the exception of Reiser, which has 
failed it dramatically.

Now, if you shut down a buffered filesystem improperly then yes, you should 
expect filesystem corruption to occur (though most of the time you will get 
lucky and be fine).  Even there, I've not had filesystem corruption problems 
with either XFS or JFS (though data can and does get lost/corrupted when the 
power is interrupted in this fashion).  ext3 the verdict is still out on 
(I've only been playing with it on one machine ... thus far no problems but 
more testing is required to be certain).

I've got GNU/Linux systems running as routers that have uptimes measured in 
hundreds of days (one of them for 460 days last I checked), with never a 
disruption or spontaneous filesystem going corrupt (they are using ext2).  
Every single reiserfs installation I had (6 or 7 IIRC) had corrupt 
filesystems that were unrecoverable within 6 months ... despite having never 
been improperly shut down or otherwise mistreated in a fashion that would 
lead one to expect, or accpet, such behavior.

Based on this experience I do not consider Reiserfs at all safe to deploy.  
XFS is safe, as long as you're not aggressively hacking the kernel (it is 
intrusive, so mucking about with other kenel hacks can affect its 
reliability.  For this reason, if you're using XFS you should stick to stock 
kernels to which only the XFS patch has been applied IMHO).  JFS also appears 
to be very safe.  Ext2 is very safe, as long as you treat it properly (do not 
shutdown improperly, and keep on a UPS if there is a concern about power 
reliability), or turn buffering off (this will slow it down, but make it safe 
even in error prone situations, such as working with unstable, experimental 
kernels or a buggy X installatino).  Ext3 appears to be ok, but I haven't 
used it enough to know that with certainty.  I tend to treat my ext3 
installation as an ext2 filesystem, so I haven't really put the journalling 
to a thorough test yet.

Reiser comes nowhere near being as safe or stable as these alternatives (with 
the possible exception of ext3 which I need to do more testing with).

Jean.

Re: [gentoo-dev] reiserfs

[Ben Lutgens]

[-- Attachment #1: Type: text/plain, Size: 1600 bytes --]

On Tue, May 14, 2002 at 11:21:02AM -0500, Jean-Michel Smith wrote:
>
>First, I've had resier lose data on systems that were running fine, i.e. were 
>NOT shut down improperly, or suffered a kernel hang, or any other sort of 
>disruption that one could reasonably expect would lead to filesystem 
>corruption.

I've seen these types of corruptions as well. It's as if the writes aren't
synced to disk in a timely manner and therefor just don't take place. The
journaling in reiser is supposed to prevent this. I've also had low memory
issues with reiser, when doing large amounts of small block IO, it'll just
randomly corrupt pieces of data. I didn't test too much to ascertain the
cause of the effect, suffice it to say that it was enough to make me stay
away from it.

>Reiser comes nowhere near being as safe or stable as these alternatives (with 
>the possible exception of ext3 which I need to do more testing with).

Stephen Tweedie will freely admit that ext2 is kinda haphazzard in the way
it does certain things but relies HEAVILY on a very good fsck. Now
reiserfsck on the otherhand (and I can't back this up either) appears to
just through out inconsistent inodes rather than trying to repair them. 

All just heresy of course, i have absolutely 0 data to back any of this up,
but I for one wouldn't use reiser in production. 

-- 
Ben Lutgens				 | http://people.sistina.com/~blutgens/	
System Administrator	 | http://www.sistina.com/
Sistina Software Inc. | 

"I got a wife and kids too but you don't see me out here stealing Imperial
Droids now do ya?"

[-- Attachment #2: Type: application/pgp-signature, Size: 240 bytes --]

Re: [gentoo-dev] reiserfs

[Alexander Gretencord]

On Tuesday 14 May 2002 17:39, Jean-Michel Smith wrote:
> Reiserfs is NOT ready for production use, and the gentoo FAQ is both wise
> and friendly for pointing that out and guiding people away from that
> particular folly.

If you have such experiences it's ok to say so _but_ you have to say why you 
don't recommend reiserfs (or give a pointer to some place where your thesis 
is backed). I've not had any bad experience with ext2 on my workstation at 
home so I'd say it's ok but at work on the latop after one crash I had to do  
a fsck'ing long fsck which then aborted and let me run the whole damn thing 
again manually. So after that I'd say ext2 sucks. But that may not be 
apparent to someone else who has only had good exp like I had at home. So he 
won't believe me if all I say is "ext2 sucks dont use it it's unstable"

> There are plenty of other, much safer filesystems to use, including XFS (if
> you don't need bleeding edge experimental features, e.g. can be happy with
> stock 2.4.18 kernel + xfs patches), JFS, ext3, ext2 (no journalling), and
> so forth.

There are problems with those too. As said earlier (and Mark mentions it too) 
ext2 sucks :) I have no personal experience with JFS but the german computer 
magazine c't tested reiser, ext3, XFS and JFS lately and they had very 
serious stability problems with JFS. So it's basically ext3 against XFS. One 
is a very intrusive patch and the other is ext2 + journaling which is fine if 
you only need a journaling fs and don't care about other deficiencies of 
ext2.

> That having been said, there is a plethora of hard evidence as well as
> anectdotal experiences to learn from, and to indicate that reiser really
> isn't a safe choice to be making.  This is reflected in the gentoo
> installation documentation, IMHO exactly as it should be.

As said earlier, it's ok to tell the people that reiser is not stable in your 
opinion, but tell them _why_! That's my whole point, nothing about being 
emotionally attached to reiser (tho I use it and am happy, but I also use XFS 
and am happy)


Alex

-- 
"They that can give up essential liberty to obtain a little temporary safety 
deserve neither liberty nor safety."
Benjamin Franklin

Re: [gentoo-dev] reiserfs

[Per Wigren]

Tuesday 14 May 2002 19.07 skrev Alexander Gretencord:
> I have no personal experience with JFS but the german computer
> magazine c't tested reiser, ext3, XFS and JFS lately and they had very
> serious stability problems with JFS. So it's basically ext3 against XFS.

JFS is the only FS I've had problems with.. The last version I used was 1.0.3 
though, so it may have gotten a bit better now...
Still, using JFS corrupted my 120GB LV full of OGGs and pr0n... :P About 1 of 
10 files became unusable.. Sometimes when opening a sourcefile on that LV it 
could look like:

printf("blablabla");
while {("&%#)"&%)#/"%"(=¤=&("#=/¤=#¤(="##¤&¤)/(?
¤#)(&#¤)("/(=)#%%"%%
#¤(/#(¤/&"#(/#")¤/"#(¤"#¤(/"#¤exit(0);

1-3KB of text were missing and replaced with junk...

Eventually I whiped the whole 120GB LV because too much was corrupt anyway, 
and reformatted it with reiserfs. That was more than a year ago. I haven't 
had a problem since...

// Wigren

Re: [gentoo-dev] reiserfs

[Matthew Kennedy]

On Tue, 2002-05-14 at 12:22, Per Wigren wrote:
> JFS is the only FS I've had problems with.. The last version I used was 1.0.3 
> though, so it may have gotten a bit better now...
> Still, using JFS corrupted my 120GB LV full of OGGs and pr0n... :P About 1 of 
> 10 files became unusable.. Sometimes when opening a sourcefile on that LV it 

Hmmm... I had a LVM on RAID-0 disaster with JFS after a week which I
never had with ext3 or XFS (using thoses for months). Although I can't
prove it, I suspect JFS somehow caused the array to become corrupted.

But JFS is pretty sweet if it works for you. It is about a zippy for
small files as reiserfs, while as fast as XFS for big sequencial
accesses.

-- 
Matthew Kennedy
Gentoo Linux Developer

Re: [gentoo-dev] reiserfs

[Jean-Michel Smith]

On Tuesday 14 May 2002 01:50 pm, Matthew Kennedy wrote:

> Hmmm... I had a LVM on RAID-0 disaster with JFS after a week which I
> never had with ext3 or XFS (using thoses for months). Although I can't
> prove it, I suspect JFS somehow caused the array to become corrupted.
>
> But JFS is pretty sweet if it works for you. It is about a zippy for
> small files as reiserfs, while as fast as XFS for big sequencial
> accesses.

It sounds like there may be an LVM issue with JFS.

Just to clarify, I should note that in all my tests, on all my systems, I have 
not at any time made use of Linux Volume Management or software RAID of any 
kind.  I have made use of hardware RAID in some instances, simple SCSI or IDE 
drives in others (and of course "poor man's RAID" in many instances, which 
basically entails a nightly, or weekly, dd of one disk to an identical 
mirror).

For this reason all of the data points (reiser's consistently unreliable 
behavior over long periods of time, ext2 and XFS's excellent behavior, and 
JFS apparently good behavior) I've provided have NOT involved interaction 
with LVM.  Indeed, I do not even compile support for LVM into the kernel as a 
rule.

I hesitate to spam a bunch of links here, but a simple google search on 
Reiserfs and data corruption in both the web and news.google.com search 
engines provide amply justification for Gentoo warning the unwary away from 
using reiser.  That having been said, if people want to add some footnotes to 
some of those threads, and people's personal, anectdotal experiences, I don't 
think that would be a bad thing at all.

Removing the warning as some advocate would be, IMHO.

Jean.

Re: [gentoo-dev] reiserfs

[Mark Bainter]

Alexander Gretencord [arutha@gmx.de] wrote:
> As said earlier, it's ok to tell the people that reiser is not stable in your 
> opinion, but tell them _why_! That's my whole point, nothing about being 
> emotionally attached to reiser (tho I use it and am happy, but I also use XFS 
> and am happy)

Just as a point of interest, I'm not attached to it either.  I'm currently
using XFS on my laptop, as I tend to keep a standard enough kernel there
that I can manage any conflicts that do come up.  I've not really had time
to research JFS.  

In general, I don't think it's necessary to warn people about using it.  If
someone is interested in using a filesystem it's their responsibility to 
read about it first and make an educated decision.  It's not the place of a
few people to tell you what you shouldn't run, with the exception of proper
labelling of experimental software.  Giving links to more information about
a filesystem is one thing, warning them away from using it seems to me a 
bad idea.

Unless of course you are shooting for the "Stupid User" segment of the
population.  However, I think there are more than enough distributions to
covering that already.  

Re: [gentoo-dev] reiserfs

[Alexander Gretencord]

On Tuesday 14 May 2002 19:49, Mark Bainter wrote:
> In general, I don't think it's necessary to warn people about using it.  If
> someone is interested in using a filesystem it's their responsibility to
> read about it first and make an educated decision.  It's not the place of a
> few people to tell you what you shouldn't run, with the exception of proper
> labelling of experimental software.  Giving links to more information about
> a filesystem is one thing, warning them away from using it seems to me a
> bad idea.

Well I think it's ok to tell people something like "I've had great problems 
with that fs, here's a link to more info about it." This would go as a kind 
reminder that there may be problems I think. But as it stands now it's worth 
nothing.

> Unless of course you are shooting for the "Stupid User" segment of the
> population.

Well you don't need to be a filesystem guru to install gentoo so I think that 
it's ok to warn about a filesystem _if_ and only if you do say why.


Alex

-- 
"They that can give up essential liberty to obtain a little temporary safety 
deserve neither liberty nor safety."
Benjamin Franklin

Re: [gentoo-dev] reiserfs

[Mark Bainter]

Alexander Gretencord [arutha@gmx.de] wrote:
> Well I think it's ok to tell people something like "I've had great problems 
> with that fs, here's a link to more info about it." This would go as a kind 
> reminder that there may be problems I think. But as it stands now it's worth 
> nothing.

Except that it still only reflects a few people's opinion.  Unless you have
the overwhelming support of the community that a particular filesystem is
bad news (for example, if it got yanked from the kernel, or the kernel
developers stuck warnings all over it because it's so dangerous, like the
ntfs write code) then you have an argument.  Even then I wouldn't do it.

The reason is that then you have to track that.  You have to watch its 
development and remember to go back and change it all in the documentation
later when it gets fixed up...or disappears from active development.

Instead, let users educate themselves, and make their own choices, and 
enable them to make those choices.

> > Unless of course you are shooting for the "Stupid User" segment of the
> > population.
> 
> Well you don't need to be a filesystem guru to install gentoo so I think that 
> it's ok to warn about a filesystem _if_ and only if you do say why.

No, and you don't need to be an expert mechanic to buy a car either.  But
if you want to pick a good one, you at least need to educate yourself a 
bit on how they perform, and do some reading.  You might get lucky by just
guessing or trusting the first salesperson you meet, but I wouldn't count
on it.

Re: [gentoo-dev] reiserfs

[Alexander Gretencord]

On Tuesday 14 May 2002 20:32, Mark Bainter wrote:
> Except that it still only reflects a few people's opinion.  Unless you have
> the overwhelming support of the community that a particular filesystem is
> bad news

Yeah right. There obviously are people that have problems with xfs (and as you 
can see JMS said there was JFS that you could use, I gave an example about it 
not being stable and some others joined in though _he_ would recommend it... 
You're probably right to leave the fs part out completely.

> The reason is that then you have to track that.  You have to watch its
> development and remember to go back and change it all in the documentation
> later when it gets fixed up...or disappears from active development.

Yeah thats too true.


Alex

-- 
"They that can give up essential liberty to obtain a little temporary safety 
deserve neither liberty nor safety."
Benjamin Franklin

Re: [gentoo-dev] reiserfs

[Mikko Moilanen]

On Tuesday 14 May 2002 17:49, Mark Bainter wrote:

> In general, I don't think it's necessary to warn people about using it.  If
> someone is interested in using a filesystem it's their responsibility to
> read about it first and make an educated decision.  It's not the place of a
> few people to tell you what you shouldn't run, with the exception of proper
> labelling of experimental software.  Giving links to more information about
> a filesystem is one thing, warning them away from using it seems to me a
> bad idea.

Its their responsibility to use it or not to use it. If somebody kindly writes 
so exact and clear documentation he can say his opinions about things, and I 
think it is  _very_ nice if he actually does so. He has right to do so. Its 
good if he does do so. My and others responsibility is to judge them and make 
choises.

Otherway nobody would document nothing more than have to at minimum and that 
would be catastrophic.

> Unless of course you are shooting for the "Stupid User" segment of the
> population.  However, I think there are more than enough distributions to
> covering that already.

Nobody knows everyting already nor nobody have time or resources to seek out.

..and your opinions were important too.
-- 

Mikko

Re: [gentoo-dev] reiserfs

[Bill Kenworthy]

Mandrake also push reiserfs and the bad experiances seem to have been
few and far between (in the early days it was /boot install issues and
unfamiliarity with a new FS).  Personally, in approx two years I have
had one minor corruption of one directory (improper shutdown when power
went in an electrical storm), but have lost two WHOLE ext2 systems in
that time, besides odd data files when power is unexpectedly lost.

The question came up on a local lug as well, with reiserfs and ext3
seeming to be the top choices, and from memory xfs seemed to be bagged
(cant remember why - something to do with the linux implementation?).

Reiserfs did get some bad press in the early days, and I think that may
be a hangover that effects peoples thinking.  I think this is a case of
YMMV, and as far as I am concerned, gentoo is the odd one out by not
reccomending reiserfs, and because there seems to be little
documentation to back it up its point of view, but a fair bit of
experiance saying reiserfs is reasonably stable.  There is also the
possibility that the problem is gentoo's implementation of reiser (that
is, other distro's patch it for some known problems).

Note that I dont regard any FS as "totally stable", but from experiance
and reccomendations, reiserfs and ext3 seem to near the top,
particularly if you NEED the protection journalling offers.  Its only
when I joined this list that I have come across people reccomending xfs.

BillK

On Wed, 2002-05-15 at 01:49, Mark Bainter wrote:
> Alexander Gretencord [arutha@gmx.de] wrote:
> > As said earlier, it's ok to tell the people that reiser is not stable in your 
> > opinion, but tell them _why_! That's my whole point, nothing about being 
> > emotionally attached to reiser (tho I use it and am happy, but I also use XFS 
> > and am happy)

Re: [gentoo-dev] reiserfs

[Jean-Michel Smith]

On Tuesday 14 May 2002 05:44 pm, Bill Kenworthy wrote:

> The question came up on a local lug as well, with reiserfs and ext3
> seeming to be the top choices, and from memory xfs seemed to be bagged
> (cant remember why - something to do with the linux implementation?).

The patch touches a great many files.  Those who want to apply other patches 
typically find the xfs patch means doing some work by hand, which puts most 
people off.  XFS is rock solid in my experience, even in the face of power 
outages, untimely shutdowns, and the like, but I am conservative and only run 
it patched against stock kernels (e.g. xfs-sources, which is 2.4.18+xfs 
patches only).  Those wanting to play with other experimental patches 
generally avoid XFS because of this (as do I on the machines I do that sort 
of thing on) for this reason.

> Reiserfs did get some bad press in the early days, and I think that may
> be a hangover that effects peoples thinking.  I think this is a case of
> YMMV, and as far as I am concerned, gentoo is the odd one out by not
> reccomending reiserfs, and because there seems to be little
> documentation to back it up its point of view, but a fair bit of
> experiance saying reiserfs is reasonably stable. 

YMMV is reason enough to not recommend a filesystem, when your milage is 
varying with respect to spontaneous filesystem corruption!  Gentoo may be the 
odd one out on this, but in my opinion that says a great deal positively 
about the technical expertise and caution of the Gentoo developers, 
particularly in light of my own experiences.

These will be my final comments on the subject.  Looking at my notes and log 
entries, I had a total of 7 machines (out of 9 total deployed) go south with 
Reiserfs on them (unrecoverable filesystem corruption, including lost 
directories, strangely null files, and in one case oddly corrupt 
files/filenames that were undeletable).  None were due to kernel oopses, 
untimely shutdowns, or any other cause that would lead one to expect 
filesystem troubles, they were all apparently spontaneous, and all happened 
within 6 months of being deployed.  The last two machines were migrated off 
of Reiser (onto ext2) before they could screw up, having been in use only 
about three months.

The corruptions happened between April and August of last year (2001).  5 
machines were running Mandrake, one Red Hat, and one Debian. (~3 months 
testing, ~8 months deployed.  I was incorrect in an earlier post when I said 
none lasted more than 6 months ... one machine survived 9 months before 
problems arose, and another didn't suffer filesystem corruption until 7 
months after deployment.  All of these machines are on 24/7)

My friend had his Suse Reiserfs go south (entire directory tree spontaneously 
vanished, but filesystem usage remained the same and even continued to grow) 
six weeks ago (April 2002), so this is by no means an early development 
glitch that is now ancient history we can comfortably dismiss and forget 
about.  (I do not know how long he had had the machine deployed for, but I 
can find out if anyone is really interested).

XFS is annoying because the patch is big, and sometimes one must wait a week 
or two after a kernel is released before a patch for xfs exists.  In the case 
of gentoo, where multiple cool patches are being applied to an experimental, 
pre-release of 2.4.19, we had to wait a week or two before some kind, 
enterprising soul managed to work the patch into the OS (testing on -r5 is 
looking very good, fwiw).  That having been said, unless one is desperate for 
a particular fix one is generally wise to wait a week or two after a kernel 
release before deploying it in a production environment anyway, so this 
(admittedly minor) irritation is mitigated for the most part.  

I've beaten on XFS under just about every condition imaginable (minus LVS, 
which I do not use) and have yet to be able to make it corrupt the 
filesystem.  I've even deliberately caused kernel oopses by trying to compile 
glibc on a kernel with high-mem, or a machine with 1 GB RAM, and been unable 
to cause damage to the filesystem.  It appears to be very solid and does not 
corrupt spontaneously.  (about 3.5 years fairly rigorous testing and very 
rigorous usage, including 2 enterprise NFS servers on large RAID devices and 
several developer workstations. Of all the filesystems I've tested, this one 
has been tested the most thoroughly, except of course for ext2 which I've 
been using a great deal longer)

JFS I've done less testing with.  It appears to be pretty good, but others 
have reported LVS corruption which may have been caused by JFS.  I haven't 
beaten on it nearly as hard as I have XFS, so I cannot say with certainty 
that it is reliable, but thus far I've yet to have it screw up.  Not exactly 
a ringing endorsement, but a cautious "it looks ok so far." (~3 months casual 
testing)

ditto ext3.  It needs more testing.  It seems to do alright thus far, but I 
tend to treat it as I would an ext2 filesystem. (~5 months casual testing).

ext2 is very solid, provided it is treated correctly (no improper shutdowns or 
power-offs), or buffering is turned off.  It does not corrupt spontaneously, 
ever.  BUT, and this is a big BUT, it can and does become corrupted if it is 
not shutdown properly (and this can happen due to system hangs, e.g. X with 
Nvidia drivers on some configurations, power outage, impatient user hitting 
the reset switch, etc.).  Most of the time it will recover through fsck, but 
not always, and I echo others who have lost ext2 filesystems that have been 
unrecoverably corrupted in this way.  It is why I prefer journalled 
filesystems and have gone to deploying XFS where possible and practical 
(often, but not always, the case due to the patch's size and complexity), and 
why I am keeping an eye on JFS and others.

It is important to note that filesystem corruption due to untimely shutdowns, 
which both ext2 and reiser have suffered from, are a completely different 
animal from the apparent spontaneous loss of data that is my major complaint 
with Reiser, and why I am so vocal in defending Gentoo's word of caution 
regarding it.

Kernel oopses:  all bets are off for any filesystem (though I've yet to be 
able to get XFS to corrupt from this, it is theoretically possible AFAICT 
since something might be going on within the kernel's vfs layer when the hang 
happens.  This is the only situation in which I find filesystem corruption in 
a journaled filesystem to be at all forgivable.

The only filesystem I have ever experienced that has corrupted itself during 
normal operations, with no unexpected reboots, kernel oopses, or other 
mitigating circumstances to explain the corruption, is Reiserfs, and these 
experiences are all within the last 14 months.

The only filesystem I've been unable to corrupt has been XFS.  (JFS and ext3 
do not count, I haven' t beaten on them the way I have ext2, XFS, and 
Reiser).

So while people should continue to experiement with Reiser (after all, that is 
how these sorts of bugs will be found and fixed), a word of caution is IMHO 
certainly in order, regardless of whether or not that makes Gentoo "the odd 
man out" or not.

Jean.

Re: [gentoo-dev] reiserfs

[Spider]

[-- Attachment #1: Type: text/plain, Size: 630 bytes --]

begin  quote
On Tue, 14 May 2002 19:10:31 -0500
Jean-Michel Smith <jsmith@kcco.com> wrote:

> 
> The only filesystem I've been unable to corrupt has been XFS.  (JFS
> and ext3 do not count, I haven' t beaten on them the way I have ext2,
> XFS, and Reiser).



I have to counter this with an anekdote from the recent chiba outage...
the XFS long scheluded writes made it wipe some files/changes and we had
to restore from backups.... so yes, XFS can combust in circumstances.

//Spider
--
begin  .signature
This is a .signature virus! Please copy me into your .signature!
See Microsoft KB Article Q265230 for more information.
end

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] reiserfs

[Jean-Michel Smith]

On Tuesday 14 May 2002 07:39 pm, Spider wrote:
> begin  quote
> On Tue, 14 May 2002 19:10:31 -0500
>
> Jean-Michel Smith <jsmith@kcco.com> wrote:
> > The only filesystem I've been unable to corrupt has been XFS.  (JFS
> > and ext3 do not count, I haven' t beaten on them the way I have ext2,
> > XFS, and Reiser).
>
> I have to counter this with an anekdote from the recent chiba outage...
> the XFS long scheluded writes made it wipe some files/changes and we had
> to restore from backups.... so yes, XFS can combust in circumstances.

That is very good to know!  Thanks!

Do you remember how you triggered the condition?

Jean.

Re: [gentoo-dev] reiserfs

[Spider]

[-- Attachment #1: Type: text/plain, Size: 1590 bytes --]

begin  quote
On Tue, 14 May 2002 19:07:20 +0200
Alexander Gretencord <arutha@gmx.de> wrote:

> 
> There are problems with those too. As said earlier (and Mark mentions
> it too) ext2 sucks :) I have no personal experience with JFS but the
> german computer magazine c't tested reiser, ext3, XFS and JFS lately
> and they had very serious stability problems with JFS. So it's
> basically ext3 against XFS. One is a very intrusive patch and the
> other is ext2 + journaling which is fine if you only need a journaling
> fs and don't care about other deficiencies of ext2.
> 
For a while (1.0.1-1.0.5 or so) I was testing and packaging JFS for
redhat-derived systems, and it was snappy and nice, but the fsck tools
needed some serious work.. and although it worked nicely in most cases
(had some issues with rejects and all that crap) when the system well
crapped up and fscked, It worked... until the last time..... then things
didn't work. at all :p  I lost directory contents on my /home (no biiig
loss since I had backups) and was generally fed up with the small (500
Meg) /home partition, so JFS died on my system then..

since that time its come further and I'm actually inclined on testing it
again, though Id need more harddrive space for that... (hint hint ;)

What I liked about JFS contra XFS was that it wasn't as intrusive as XFS
.. it was a "nice" patch that touched fewer files and modified less of
the kernel behaviour.


//Spider

--
begin  .signature
This is a .signature virus! Please copy me into your .signature!
See Microsoft KB Article Q265230 for more information.
end

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

RE: [gentoo-dev] reiserfs

[Sean Mitchell]

> First, an appeal to authority (Suse in this case) is a 
> logical fallacy you 
> should not engage in.  Just because Suse ships reiser with 
> their distribution 
> doesn't make it stable or safe for production use.

[rest deleted]

These are good points. I suggest that someone put these details in the FAQ
which doesn't say much other than "we don't recommend it".

Sean

 

[gentoo-dev] reiserfs

[Brady Wied]

I was reading your FAQ on your website talking about reiserfs being
unstable with 2.4 kernels.  I have run it since 2.4.5 on slackware,
using it on 2.4.5, 2.4.12, and 2.4.18 now with no problems. What
problems did you guys experience?

Brady

Re: [gentoo-dev] reiserfs

[Spider]

[-- Attachment #1: Type: text/plain, Size: 567 bytes --]

begin  quote
On Tue, 14 May 2002 09:44:39 -0500
"Brady Wied" <wied@shs.tamu.edu> wrote:

> I was reading your FAQ on your website talking about reiserfs being
> unstable with 2.4 kernels.  I have run it since 2.4.5 on slackware,
> using it on 2.4.5, 2.4.12, and 2.4.18 now with no problems. What
> problems did you guys experience?
> 
Files being filled with all NULL characters after reboot / unclean
remount.

//Spider

--
begin  .signature
This is a .signature virus! Please copy me into your .signature!
See Microsoft KB Article Q265230 for more information.
end

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: [gentoo-dev] reiserfs

[Alexander Gretencord]

On Tuesday 14 May 2002 23:17, Spider wrote:
> Files being filled with all NULL characters after reboot / unclean
> remount.

Well I've had similar things happen. Actually my files did contain something, 
my xmms.m3u had a shell script I recently edited in it etc. ... But this is 
due to reiser only journaling metadata. This can happen with XFS and JFS too 
(ext3 can also journal data).


Alex

-- 
"They that can give up essential liberty to obtain a little temporary safety 
deserve neither liberty nor safety."
Benjamin Franklin