Gentoo Forums
Gentoo Forums
Gentoo Forums
Quick Search: in
Filesystems comparison for present time (r4,r3,jfs,xfs,ext3)
View unanswered posts
View posts from last 24 hours

Goto page Previous  1, 2, 3, 4, 5  
Reply to topic    Gentoo Forums Forum Index Off the Wall
View previous topic :: View next topic  

Please READ the tables,next for which fs You want to switch ? :)
ext3
26%
 26%  [ 68 ]
jfs
10%
 10%  [ 26 ]
reiser3
17%
 17%  [ 44 ]
reiser4
35%
 35%  [ 91 ]
xfs
9%
 9%  [ 25 ]
Total Votes : 254

Author Message
i92guboj
Moderator
Moderator


Joined: 30 Nov 2004
Posts: 9810
Location: Córdoba (Spain)

PostPosted: Tue Apr 19, 2005 5:44 pm    Post subject: Reply with quote

fallow wrote:
to be sincere . Wen Im doing copying of my portage dir ? . I neved did it. hmm maybe some times when Im doing a backup.

emerge sync isnt a copy of portage dir ;)

and as I told , my /usr/src and portage is in reiser3 ;) absolutely no reason for my to user reiser4 instead of reiser3. rest of the system.

my golas for choosing fs are : interactivity of the system in fs load situations and the second is performance , not first. stablility is very individual question for each one of us.

for me isnt very matter about copying of my dir will be done in 35 or 43 seconds. I want to have very good overall interactivity of my destkop in every situtation ;)

cheers.

That's mine too.

By the way, I remember a post that I put some time ago. It was something about some machines reporting full cpu usage when doing disk I/O. I'm referring to modem pc's which are correctly configured and with dma settings enabled. Total interactivity mess, to sum up.

I saw the table above (the one that fellow posted with the output of bonnie++) and saw that ext3 is also a big cpu hog (I did not know of that at all). I'm now wondering if making a fs change in these machines will make any difference.

Of course, regardless of the opinions and the answers I get here I'll try ;) but I'm wondering if someone experiences anything similar.

Oh! Another question. About cpu, the thing is that journaled fs's will always waste more cpu cycles than non journalised ones. So i'm wondering if really ext2 would be an interactity booster, jfs is also less cpu angry, but is journalised, so I think it would be the intermediate point, but the tests show that is the slowest one, why?

And also, now that I'm asking so many question I'll ask another one now: anyone has similar test results showing also something with fat32 vs. ext3 or any other fs's? I ask this because when comes to desktop interactivity windows seems much responsive in many situations, and I wonder if the simplistic, arcaic, shit-fs has something to do with this.
_________________
Gentoo Handbook | My website
Back to top
View user's profile Send private message
yardbird
l33t
l33t


Joined: 20 Apr 2002
Posts: 688
Location: nl.leiden

PostPosted: Tue Apr 19, 2005 7:22 pm    Post subject: Reply with quote

6thpink wrote:
I agree with you that, in concept, the idea which suposedly is reiser4 is a good thing. But I truly think that they have not reached the level of atomicity or security they wanted, still. I know that I may appear to be a bit rude when talking about reiser4, but have spend so much awfull evening trying to repair fs-massacres.

It seems like the we all have our (different) experiences regarding fses :D Personally I've been using r4 on my laptop's root partition since last June (when the fs was still in heavy development), and, though there were some glitches in the beginning, I haven't lost a single bit of my data (BTW I'm using r4 also on my desktop, on a raid-0 setup - never lost data even there). I'm not saying that r4 is stable because it has worked ok for me, I'm just saying that it's extremely difficult to define such thing as "stability". A peculiar problem with r4 was that in the beginning there were many non-official third party patches which were badly broken (because of non-trivial port attempts from -mm). Also, to define "stability", one has to define the context: you cannot have a stable fs when other parts of your system aren't. I remember a single case of fs corruption here on the forums in which the fs was blamed, only to discover that the guy was using hdparm's -W (write caching) option, which is absolutely _dangerous_ for a journalled fs in case of power interruption.

6thpink wrote:
Maybe, the point here is the maturity of the code. Really, do you think that reiser4 is ready for a production system yet? I don't think so, it is still far from perfect.

Yes, I agree with you here. But please remember that r4 is not even in vanilla yet! It has not received the amount of testing the other fses have had.

6thpink wrote:
And also, now that I'm asking so many question I'll ask another one now: anyone has similar test results showing also something with fat32 vs. ext3 or any other fs's? I ask this because when comes to desktop interactivity windows seems much responsive in many situations, and I wonder if the simplistic, arcaic, shit-fs has something to do with this.

I think it has more to do with windows' memory management. I've always had the impression that it does not free memory when you close a program, so when you start a program for the 2nd time it comes up almost instantly.

6thpink wrote:
Btw, u like pink floyd? (say for your ubication info, if thats the case, here I agree with you :D )

Oh yeah! I'm a jazz person to be honest, but I'm totally hooked on Pink Floyd :D Are you the 6th one? :wink:
_________________
Albert Einstein wrote:
I consider it [...] urgently necessary for [...] workers to get together, both to protect their own economic status and [...] to secure their influence in the political field.


http://www.bluescarni.info
Back to top
View user's profile Send private message
i92guboj
Moderator
Moderator


Joined: 30 Nov 2004
Posts: 9810
Location: Córdoba (Spain)

PostPosted: Tue Apr 19, 2005 7:47 pm    Post subject: Reply with quote

yardbird wrote:
It seems like the we all have our (different) experiences regarding fses :D


Yep, that's true.

yardbird wrote:
I'm not saying that r4 is stable because it has worked ok for me, I'm just saying that it's extremely difficult to define such thing as "stability". A peculiar problem with r4 was that in the beginning there were many non-official third party patches which were badly broken (because of non-trivial port attempts from -mm). Also, to define "stability", one has to define the context: you cannot have a stable fs when other parts of your system aren't. I remember a single case of fs corruption here on the forums in which the fs was blamed, only to discover that the guy was using hdparm's -W (write caching) option, which is absolutely _dangerous_ for a journalled fs in case of power interruption.


Well, I know that "stability" is a hard thing to define. I personally consider something stable when it works reasonably well and without any glitches in almost all the imaginable system and configurations. I know that in this terms there are few things that are really stable (and I also include here the rest of filesystems), but, as I said before, in this cases I like to trust statistics, and they say that, for example, ext3 or reiser 3.6 are nearer to my concept of stability than reiser4 do. Of course this could change with the time, in fact I haven't tried reiser4 since a few months, so, maybe the thing has changed.

Anyway, the high cpu consumption is a thing that will keep me far from that filesystem. I'm also thinking about changing from ext3 to any other thing, since I maintain some machines that can't stand such high cpu activity just for the minimum disk activity and ext3 has also a so high cpu consumption.

yardbird wrote:
...remember that r4 is not even in vanilla yet! It has not received the amount of testing the other fses have had.


I know. Which bothers me about reiser4 (and a bit less but also 3.6) is its high cpu usage on most circumstances. I think that 98% cpu usage is so much, even for modern machines and high load. So, what's the point in using udma drives if your filesystem sucks a so high cpu cycles amount?

For me interactivity is important. Of course would be nice to have a filesystem capable of both: throughput capacity and light cpu usage, but, for now, this is just an utopia.

yardbird wrote:
6thpink wrote:
And also, now that I'm asking so many question I'll ask another one now: anyone has similar test results showing also something with fat32 vs. ext3 or any other fs's? I ask this because when comes to desktop interactivity windows seems much responsive in many situations, and I wonder if the simplistic, arcaic, shit-fs has something to do with this.

I think it has more to do with windows' memory management. I've always had the impression that it does not free memory when you close a program, so when you start a program for the 2nd time it comes up almost instantly.


I know that windows preloads so many things. But here there are some things that are not explanable.

Kde, for example, also preloads konqueror if you configure it that way, but even then there are some cases where konqueror will take more time to load that explorer in windows. Anyway this is not a thing to bother about (to wait 1 second is not a bad thing), I'm talking about general interactivity of the onscreen interfaces. About menus, buttons, about window switching.

I dont use kde really most of time, so don't blame it. I ussually work with fvwm wich is the fastest window manager that I have ever tried.

And about windows preloading, another thing: yes, it preloads, but it is also true that windows dont fully share all the dll's code in memory like linux do. So it has to reload many dll's even when they are still in memory, linux don't usually has to do that. So, where's the point? I really don't think that all the story is about preloading or not preloading. It's evident that a non-journaled fs is faster than a journaled one.

yardbird wrote:
6thpink wrote:
Btw, u like pink floyd? (say for your ubication info, if thats the case, here I agree with you :D )

Oh yeah! I'm a jazz person to be honest, but I'm totally hooked on Pink Floyd :D Are you the 6th one? :wink:

Ouh yeah! Here an agreement! From Luis Bonfa to Jaco Pastorius, from Aretha Franklin to Stanley Clarke. Jazz rules. And Pink Floyd also! :wink: Btw, I'm not the 6th pink really, but also have a Strat USA. :D
_________________
Gentoo Handbook | My website
Back to top
View user's profile Send private message
superstoned
Guru
Guru


Joined: 17 Dec 2004
Posts: 432

PostPosted: Tue Apr 19, 2005 8:25 pm    Post subject: Reply with quote

hey, about interactivity - isn't it that it does not matter how MUCH cpu a fs uses, but more how long codepath's are, and how interuptable the code is (eg is it pre-emption-friendly)?

so a fs might use 90 % cpu, but if some other process needs cpu, and it can get it immediately because of easy pre-emption of fs code, it is no problem...

and yes, I know, reiser4 is extremely bad here - not only high cpu usage, but also LONG codepaths (why does it need 8k stacks? duh...), badly pre-emptable. But work is being done on this.

yes, if you want a fast, responsible desktop, reiser4 is bad for you. but its getting better.

if you want the max throughput (still not sure about this, but I'm testing this) reiser4 is best. so - if you do most emerges when you sleep, put reiser4 on /usr/portage and /var/tmp/portage. I will have /usr/src and /var/tmp/portage and /usr/portage on r4, if my next test shows it is faster.

I'm going to test emerge --sync, emerge qtparted, the make of a kernel, extraction of a kernel tarball and removing of a kernel tree.
Back to top
View user's profile Send private message
nmcsween
Guru
Guru


Joined: 12 Nov 2003
Posts: 381

PostPosted: Tue Apr 19, 2005 9:56 pm    Post subject: Reply with quote

mount xfs with the following and also make a larger journal:
Code:
mount -o noatime,nodiratime,logbufs=8

_________________
Great Resources
Back to top
View user's profile Send private message
IvanYosifov
l33t
l33t


Joined: 15 Oct 2004
Posts: 778
Location: Bulgaria

PostPosted: Wed Apr 20, 2005 5:58 am    Post subject: Reply with quote

Quote:

I ask this because when comes to desktop interactivity windows seems much responsive in many situations

In my experience with Windows it puts responsiveness far above evrything else. While a desktop system should be responsive it also nice if can do some work. I guess compiling OOo is not considered by M$ a typicall workload :D Under load Windows comes to a point were all you can do is smoothly move windows around, but if there is any computationally or IO intesive task running it will take longer and longer. The system will be crawling though it will not look so. You get precieved performance at the expense of real performace.

Quote:

It's evident that a non-journaled fs is faster than a journaled one.

Of course, yet I value my data and don't have a power backup so ( at least I ) don't see it as an option ( at least not for / , maybe good for /var/tmp/portage though ).

I finally got around to running bonnie++ on my machine. The fs is reiser3. The results are:
Code:

Version 1.93c       ------Sequential Output------ --Sequential Input- --Random-
Concurrency   1     -Per Chr- --Block-- -Rewrite- -Per Chr- --Block-- --Seeks--
Machine        Size K/sec %CP K/sec %CP K/sec %CP K/sec %CP K/sec %CP  /sec %CP
home.yosifov.net 2G   227  99 34469  14 16695   5  1465  97 40912   8 174.2   3
Latency               134ms    1206ms     131ms   46729us   39045us     302ms
Version 1.93c       ------Sequential Create------ --------Random Create--------
home.yosifov.net    -Create-- --Read--- -Delete-- -Create-- --Read--- -Delete--
              files  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP  /sec %CP
                 16 18910  85 +++++ +++ 18318  99 19266  89 +++++ +++ 17455  99
Latency               101us    2415us    2485us     118us      25us     125us

They look consistent with what has been reported so far.

Looks like reiser has high CPU only on per-char operations, which are devastating for performance anyway. On block IO CPU gets as high as 14%, and that's it.

It is my understanding that DMA is about block transfers, so it can't help on byte-by-byte operation. I know little about DMA, don't kill me if wrong. This theory looks consistent with fallow's results, where all filesystems use 100% CPU on per-char operations

EDIT:
And only jfs,xfs are marginally better at sequential/random creation.
Back to top
View user's profile Send private message
yardbird
l33t
l33t


Joined: 20 Apr 2002
Posts: 688
Location: nl.leiden

PostPosted: Wed Apr 20, 2005 12:49 pm    Post subject: Reply with quote

6thpink wrote:
I know that windows preloads so many things. But here there are some things that are not explanable.

Kde, for example, also preloads konqueror if you configure it that way, but even then there are some cases where konqueror will take more time to load that explorer in windows. Anyway this is not a thing to bother about (to wait 1 second is not a bad thing), I'm talking about general interactivity of the onscreen interfaces. About menus, buttons, about window switching.

I dont use kde really most of time, so don't blame it. I ussually work with fvwm wich is the fastest window manager that I have ever tried.


Well, that kind of interactivity is in the X server's domain. Xorg/Xfree is _not_ slow compared to windows, it just seems slow ;) What I mean here is that windows uses many "tricks" to look fast, like double buffering for example (which avoids scrolling and redraw effects). Xorg is getting there too, and some widgets (like QT4 with Arthur and Gtk+ with Cairo) will implement a lot of nice features on their own. But I agree that _today_ Xorg seems to be slower than windows.

On the preloading matter I think I did not express my view clearly. I meant to say that when an application is loaded in windows it stays there until memory is reclaimed by some other process because it is running out. On linux the standard behaviour is to free that space in a more "dynamical" fashion. But we would need a memory-management expert here to settle down the matter :)

I don't think that it is FAT32 that makes windows snappy. I did not notice any difference in interactivity switching back and forth NTFS and FAT32, and NTFS is certainly heavier than FAT32 (both in features and CPU consumption). I think that under linux the difference in interactivity is more due to how fs, CPU-scheduler and IO-scheduler interact than to just the filesystem. For example my personal experience is that reiser4 is snappier with the "vanilla" O1 scheduler than with staircase, for example. OTOH reiser3+staircase make a good combination for interactivity. I guess these are the joys and woes of freedom of choice :wink:

6thpink wrote:
Ouh yeah! Here an agreement! From Luis Bonfa to Jaco Pastorius, from Aretha Franklin to Stanley Clarke. Jazz rules. And Pink Floyd also! :wink: Btw, I'm not the 6th pink really, but also have a Strat USA.

I fully agree with you on this :D I'm particularly fond of saxophonist (being one myself), and I _love_ Miles Davis. But we are going terribly OT here... when will there be a music sub-forum in OTW? :wink:
_________________
Albert Einstein wrote:
I consider it [...] urgently necessary for [...] workers to get together, both to protect their own economic status and [...] to secure their influence in the political field.


http://www.bluescarni.info
Back to top
View user's profile Send private message
superstoned
Guru
Guru


Joined: 17 Dec 2004
Posts: 432

PostPosted: Wed Apr 20, 2005 5:21 pm    Post subject: Reply with quote

yardbird wrote:
6thpink wrote:
I know that windows preloads so many things. But here there are some things that are not explanable.

Kde, for example, also preloads konqueror if you configure it that way, but even then there are some cases where konqueror will take more time to load that explorer in windows. Anyway this is not a thing to bother about (to wait 1 second is not a bad thing), I'm talking about general interactivity of the onscreen interfaces. About menus, buttons, about window switching.

I dont use kde really most of time, so don't blame it. I ussually work with fvwm wich is the fastest window manager that I have ever tried.

Well, that kind of interactivity is in the X server's domain. Xorg/Xfree is _not_ slow compared to windows, it just seems slow ;) What I mean here is that windows uses many "tricks" to look fast, like double buffering for example (which avoids scrolling and redraw effects). Xorg is getting there too, and some widgets (like QT4 with Arthur and Gtk+ with Cairo) will implement a lot of nice features on their own. But I agree that _today_ Xorg seems to be slower than windows.

well, I think X/Xorg IS slow. At least that's what con says, and I've read some others say it too - so I guess it's true. and it does not fee that fast, too - try moving a window around, see the cpu usage go up. You can easilly reach 100% - that's insane. try it under windows - no problem. Bad chaching/buffering of data, I guess. and, to quote Con Kolivas, "Once X code improves to stop burning cpu cycles when the gpu is busy (bad design) then interactive mode on will no longer be required.".
yardbird wrote:
I don't think that it is FAT32 that makes windows snappy. I did not notice any difference in interactivity switching back and forth NTFS and FAT32, and NTFS is certainly heavier than FAT32 (both in features and CPU consumption). I think that under linux the difference in interactivity is more due to how fs, CPU-scheduler and IO-scheduler interact than to just the filesystem. For example my personal experience is that reiser4 is snappier with the "vanilla" O1 scheduler than with staircase, for example. OTOH reiser3+staircase make a good combination for interactivity. I guess these are the joys and woes of freedom of choice :wink:

Again, as I said before, I don't think its the cpu usage that matters. a kernel compile uses 100% cpu, does that make your system unusable? its about how easy the code is interupted, to be able to run other code, that matters for interactivity.
Back to top
View user's profile Send private message
Sith_Happens
Veteran
Veteran


Joined: 15 Dec 2004
Posts: 1807
Location: The University of Maryland at College Park

PostPosted: Mon Apr 25, 2005 6:36 pm    Post subject: Reply with quote

Hey guys, check out my post here and see what you think. I'm trying to compare space efficiency between various filesystem types.
_________________
"That question was less stupid; though you asked it in a profoundly stupid way."
I'm the brains behind Jackass! | Tutorials: Shorewall
Back to top
View user's profile Send private message
nmcsween
Guru
Guru


Joined: 12 Nov 2003
Posts: 381

PostPosted: Mon Apr 25, 2005 11:04 pm    Post subject: Reply with quote

If your trying to optimize for size then set the block size to 512b. That will kill performance but save space.
_________________
Great Resources
Back to top
View user's profile Send private message
nmcsween
Guru
Guru


Joined: 12 Nov 2003
Posts: 381

PostPosted: Tue Apr 26, 2005 11:14 pm    Post subject: Reply with quote

Fallow could you test out the following for XFS:
Code:
mkfs.xfs -d agsize=3g -l size=64m
Code:
async,noatime,rw,logbufs=8,logbsize=32768,osyncisdsync

_________________
Great Resources
Back to top
View user's profile Send private message
erikm
l33t
l33t


Joined: 08 Feb 2005
Posts: 631

PostPosted: Fri Jun 03, 2005 6:05 pm    Post subject: Reply with quote

I have always used ext3, and I am very happy with it. I've seen some peculiarities after more or less constant NFS writes >= 1 GB for three months straight, but they were automagically fixed after a remount...
However, I did suffer some severe data loss with ext3 once:

/me working late (again)
/me has a minor caffeine deficiency (as in, just below actually testing positive for methamphetamine)
/me moves some important files around partitions without backups (after all, /methinks /me so l33t /me never makes mistakes :roll: )
/me forgets a partition mounted, rewrites the partition table, and formats the correct partition... correct in the new table, that is. :oops:

Massive data loss, and a bitch and a half to recover since the freshly formatted partition was also mounted, and thus written to.

The morale is, performance means squat when you have your head up your ass... :D

Or, I guess, never stop feeding that ten-cups-a-day habit. :wink:
Back to top
View user's profile Send private message
cbr
Apprentice
Apprentice


Joined: 05 Jan 2004
Posts: 285
Location: Tallinn/Rakvere, Estonia

PostPosted: Fri Jun 03, 2005 8:11 pm    Post subject: Reply with quote

When i have lots of mp3 files that are usually 3-8MB, is it advisable to use XFS? Do I get any more speed? Is ~5MB considered a big file in FS speech? Will XFS perform well with files like that or are they too small to unleash the real power of XFS?
Back to top
View user's profile Send private message
IvanYosifov
l33t
l33t


Joined: 15 Oct 2004
Posts: 778
Location: Bulgaria

PostPosted: Sat Jun 04, 2005 10:20 am    Post subject: Reply with quote

5MB is a normal file. Several GB is a big file. I guess all FS will handle 5MB files just the same. Don't bother, if playback is skipping or something, the problem is almost certainly elsewhere.
Back to top
View user's profile Send private message
vincent_bachelier
Tux's lil' helper
Tux's lil' helper


Joined: 26 Sep 2004
Posts: 96

PostPosted: Wed Aug 10, 2005 12:34 pm    Post subject: Reiser4 result Reply with quote

Hi,
I have move all my system on reiser4 fs
my system where on ext3 var / /var /usr /home
and ext2 for /usr/portage

Result:

emerge are 5x faster
some strange pic of cpu consuption which block computer during working ... worst for desktop use
bad result with "find" and bash-completion. all block during searching ...
gentoo don't check system even system crash ... fsck -A -a don't work

this last worst think have a really bad concequence.
after my computer have low battery, acpi shutting down it.
when I connect again my computer, a lot's of data where lost, and as gentoo doesn't check anythink, I do ls /home and lot's of IO error when said by system
I run a live cd, and do fsck.reiser4 for checking, need to rebuild-fs, and then all my document where lost !

I discourage anybody to use it now, it's the worst system I ever test ! bad gestion on gentoo for now, lot's of error after crashing (more than ext2 !)
need to often rebuild-fs with lost of file after that.

I will check xfs or jfs, I don't know the best, I want a good successor of ext3, for the moment, nothing good with reiser (3 or 4)

tell me your impression
mine is awfull

I have test all this with gentoo-sources + reiser4 patch for vanilla

ok, perhaps it work fine with vanilla + reiser4, but there my result

see ya
_________________
Shaolan Oji
Blog : http://ilwg.blogspot.com [ Gentoo Fan Blog Site ]
Project:
vbwallpapers: http://vbwallpapers.berlios.de [ get random wallpapers for your desktop ]
Back to top
View user's profile Send private message
bollucks
l33t
l33t


Joined: 27 Oct 2004
Posts: 606

PostPosted: Wed Aug 10, 2005 12:56 pm    Post subject: Reply with quote

Stick with ext3 and add the dir_index and journal=writeback options.
Back to top
View user's profile Send private message
crazycat
l33t
l33t


Joined: 26 Aug 2003
Posts: 838
Location: Hamburg, Germany

PostPosted: Wed Aug 10, 2005 1:02 pm    Post subject: Reply with quote

I use nitro sources with reiser4 and use it succesfully for over a month. You probably have got old reiser4 patches. I had problems like yours some months ago. But i tried reiser4 recently and found it stable.
Back to top
View user's profile Send private message
vincent_bachelier
Tux's lil' helper
Tux's lil' helper


Joined: 26 Sep 2004
Posts: 96

PostPosted: Wed Aug 10, 2005 1:08 pm    Post subject: reiser4 patch Reply with quote

Ok, gentoo-source + reiser4 patch (just out on namesys) is a really bad combination !
you probably haven't got computer who shutdown with some issue and replay journal are not really good ...

I have lose all my data, I'm really angry against it, thx god I have make a big backup few days ago

what performance I win with ext3 option ? writeback and dir_index ?

nitro source work probably well with reiser4 now ... but I think I will wait for including in vanilla official source ...

what about xfs and jfs ?

I'm searching the best performance for my desktop, with good multitache fs, and without any possibility of losing any data !

have some good suggestion or anythink else ?

thx for support
_________________
Shaolan Oji
Blog : http://ilwg.blogspot.com [ Gentoo Fan Blog Site ]
Project:
vbwallpapers: http://vbwallpapers.berlios.de [ get random wallpapers for your desktop ]
Back to top
View user's profile Send private message
nxsty
Veteran
Veteran


Joined: 23 Jun 2004
Posts: 1556
Location: .se

PostPosted: Wed Aug 10, 2005 1:42 pm    Post subject: Re: reiser4 patch Reply with quote

vincent_bachelier wrote:

what about xfs and jfs ?

I'm searching the best performance for my desktop, with good multitache fs, and without any possibility of losing any data !


Both XFS och ReiserFS v3 have very good performance. ReiserFS v3 is particular good at handling a lot of small files like the portage tree, but it also uses slightly more CPU than XFS and EXT3. Search google for some benchmarks!
Back to top
View user's profile Send private message
fallow
Bodhisattva
Bodhisattva


Joined: 08 Jan 2004
Posts: 2206
Location: Poland

PostPosted: Wed Aug 10, 2005 1:52 pm    Post subject: Reply with quote

OMG another one thread about the same ? ;)

check the big one here -> http://forums.gentoo.org/viewtopic-p-2301717.html#2301717 :P

btw. Im on ext3 with dir_index and mounted with journal=writeback
there is nothing better for me at the moment.
reiser* is bad at the start because of too high cpu usage and latencies ( especially reiser4 ).
xfs and jfs are not _native_ for linux, even the drivers are good for SGI or AIX , the linux drivers for xfs/jfs its another thing.

cheers.
_________________
"Time is a companion that goes with us on a journey. It reminds us to cherish each moment, because it will never come again. What we leave behind is not as important as how we have lived" J-L. Picard ;)
Back to top
View user's profile Send private message
Deathwing00
Moderator
Moderator


Joined: 13 Jun 2003
Posts: 4078
Location: Barcelona, Spain

PostPosted: Wed Aug 10, 2005 4:53 pm    Post subject: Reply with quote

"Reiser4 result" merged to "Filesystems comparison for present time (r4,r3,jfs,xfs,ext3)"
Back to top
View user's profile Send private message
diabx0r
Apprentice
Apprentice


Joined: 08 Jan 2004
Posts: 207

PostPosted: Tue Sep 12, 2006 4:42 pm    Post subject: Reiser VS xfs VS ext3... HELP! :D Reply with quote

Well, I am about to setup my filesystems and I'm curious on the pros and cons of these filesystems.

I trust you guys more than ANYONE. :)


I'm sure they're all fine, but I really want to know what you all prefer and why.

Thanks for your time.

-Alex
Back to top
View user's profile Send private message
mdeininger
Veteran
Veteran


Joined: 15 Jun 2005
Posts: 1740
Location: Emerald Isles, overlooking Dublin's docklands

PostPosted: Tue Sep 12, 2006 5:31 pm    Post subject: Reply with quote

i think this thread should answer your question:
http://forums.gentoo.org/viewtopic-t-490049-highlight-filesystem.html
_________________
"Confident, lazy, cocky, dead." -- Felix Jongleur, Otherland

( Twitter | Blog | GitHub )
Back to top
View user's profile Send private message
nixnut
Bodhisattva
Bodhisattva


Joined: 09 Apr 2004
Posts: 10974
Location: the dutch mountains

PostPosted: Tue Sep 12, 2006 5:35 pm    Post subject: Reply with quote

merged above post here
_________________
Please add [solved] to the initial post's subject line if you feel your problem is resolved. Help answer the unanswered

talk is cheap. supply exceeds demand
Back to top
View user's profile Send private message
dentharg
Guru
Guru


Joined: 10 Aug 2004
Posts: 438
Location: /poland/wroclaw

PostPosted: Wed Oct 04, 2006 7:49 am    Post subject: Reply with quote

I've been an XFS user for a looong time.. until recent power failure and resulting data corruption. XFS does only metadata journaling and not data journaling (which can be enabled in ext3). Each dirty reset of your machine can result in trash being put into opened files. Not so good.

I've switched to ext3. Happy now.
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Gentoo Forums Forum Index Off the Wall All times are GMT
Goto page Previous  1, 2, 3, 4, 5
Page 5 of 5

 
Jump to:  
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum