View previous topic :: View next topic |
Do you need Reiser4? |
It would be great to have it with gentoo-dev-sources and/or gentoo-sources |
|
65% |
[ 367 ] |
It is enough if it works with other sources. |
|
15% |
[ 88 ] |
I don't need Reiser4 at all, so I don't care. |
|
18% |
[ 103 ] |
What is Reiser4? |
|
0% |
[ 2 ] |
|
Total Votes : 560 |
|
Author |
Message |
rzZzn Tux's lil' helper
Joined: 24 Aug 2004 Posts: 96 Location: Sweden
|
Posted: Fri Oct 14, 2005 9:18 pm Post subject: |
|
|
electrofreak wrote: | does it work on amd64 as of yet? |
I use it on my amd64 box
I have it on everything save /boot and some old pata discs that I still use.
Haven't had any lost data at all. |
|
Back to top |
|
|
Jake Veteran
Joined: 31 Jul 2003 Posts: 1132
|
Posted: Fri Oct 14, 2005 10:54 pm Post subject: |
|
|
electrofreak wrote: | does it work on amd64 as of yet? |
It's worked for me ever since they fixed the reiser4_find_next_zero_bit bug, and I'm the one who discovered a reliable way to get a panic or hardlock in about half an hour with that bug. For some reason I'm good at breaking filesystems . If the developers weren't so busy, I'd be complaining about PPC and reporting bugs. |
|
Back to top |
|
|
ciaranm Retired Dev
Joined: 19 Jul 2003 Posts: 1719 Location: In Hiding
|
Posted: Fri Oct 14, 2005 11:02 pm Post subject: |
|
|
electrofreak wrote: | does it work on amd64 as of yet? |
Heh. It doesn't even work on x86 yet. |
|
Back to top |
|
|
electrofreak l33t
Joined: 30 Jun 2004 Posts: 713 Location: Ohio, USA
|
Posted: Fri Oct 14, 2005 11:08 pm Post subject: |
|
|
ciaranm wrote: | electrofreak wrote: | does it work on amd64 as of yet? |
Heh. It doesn't even work on x86 yet. |
Lol. I've heard of many people who haven't had a problem with it really. I myself have yet to use it. But next time I feel the need to reinstall gentoo on my laptop (due to too much cruft and it's easier to just reinstall sometimes) I plan to use reiser4. But that probably wont be for some time though. _________________ Desktop: ABit AN8, Athlon64 X2 4400+ 939 2.75GHz, 2x1GB Corsair XMS DDR400, 2x160GB SATA RAID-0, 2x20"W, Vista Ultimate x64
Laptop: 15.4" MacBook Pro 2.4Ghz, 2x1GB RAM, 160GB, Mac OS X 10.5.1
Server: PIII 550Mhz, 3x128MB RAM, 160GB, Ubuntu Server 7.10 |
|
Back to top |
|
|
i92guboj Bodhisattva
Joined: 30 Nov 2004 Posts: 10315 Location: Córdoba (Spain)
|
Posted: Fri Oct 14, 2005 11:47 pm Post subject: |
|
|
electrofreak wrote: | ciaranm wrote: | electrofreak wrote: | does it work on amd64 as of yet? |
Heh. It doesn't even work on x86 yet. |
Lol. I've heard of many people who haven't had a problem with it really. |
Again, there are people who drops from a third flat and survive with no harm. That does not mean that droping yourself from a third flat is right. Though of course, if you want, you are allowed to. |
|
Back to top |
|
|
superstoned Guru
Joined: 17 Dec 2004 Posts: 432
|
Posted: Sat Oct 15, 2005 9:15 am Post subject: |
|
|
6thpink wrote: | electrofreak wrote: | ciaranm wrote: | electrofreak wrote: | does it work on amd64 as of yet? |
Heh. It doesn't even work on x86 yet. |
Lol. I've heard of many people who haven't had a problem with it really. |
Again, there are people who drops from a third flat and survive with no harm. That does not mean that droping yourself from a third flat is right. Though of course, if you want, you are allowed to. | still, there are lots of ppl that drop themselves from a flat. there must be something to it, really... gonna try it
i think you guys overexagerate the dangers of running reiser4. it's been out for quite some time now, lots of bugs have been smashed. ok, it might be a litle more dangerous compared to ext2/3 but no filesystem is safe, you have to make backups anyway. the main reason i don't use it is its latency, not the danger involved... i've lost data with ext3, xfs and reiserfs, and since these are supposed to be safe, i don't trust filesystems at all.
i don't understand the 'reiser4 sucks' attitude some ppl have. are you so enthousiastic about all new things? i suppose you run linux 1.0? you hate gnome > 1.4? no openoffice beta's, of course? sorry, but i think more ppl lose data by using unstable software like openoffice beta's...
are you afraid for the bugreports? is it so hard to ignore them, or will there be so many of them? if its the bugreports, you could say "try reiser4, its cool, but don't send bugreports as we can't/won't fix it"...
i don't understand the negative attitude, and as such it seems a bit childish. to me.
reiser4 might not be very stable, as it is new. but it's a fcking filesystem, why all the fuzz? let ppl try it, why not, choice is one of the nice things FOSS has to offer?!? |
|
Back to top |
|
|
djm Arch/Herd Tester
Joined: 12 Apr 2004 Posts: 690 Location: Wadham College, Oxford
|
Posted: Sat Oct 15, 2005 11:47 am Post subject: |
|
|
New things are great. Poorly designed and implemented new things aren't. I'd really like not reiser4 not suck - it's seriously fast and efficient in terms of disk space usage (but then there is the latency issue...)
People who are actually knowledgeable enough to create useful bug reports when their filesystem goes tits up should try reiser4 if they want to. People who use reiser4 because they think it's cool but don't have a clue what to do when things go wrong (other than whine about it) are probably doing more harm than good. _________________ the forums.gentoo.org poster formally known as metal leper |
|
Back to top |
|
|
superstoned Guru
Joined: 17 Dec 2004 Posts: 432
|
Posted: Sat Oct 15, 2005 12:30 pm Post subject: |
|
|
djm wrote: | New things are great. Poorly designed and implemented new things aren't. I'd really like not reiser4 not suck - it's seriously fast and efficient in terms of disk space usage (but then there is the latency issue...)
People who are actually knowledgeable enough to create useful bug reports when their filesystem goes tits up should try reiser4 if they want to. People who use reiser4 because they think it's cool but don't have a clue what to do when things go wrong (other than whine about it) are probably doing more harm than good. | amen. i'm no coder so i can't say reiser4 has a bad design. some say it has, some say it is great. well, i can only say how it works - and the latency kills me |
|
Back to top |
|
|
i92guboj Bodhisattva
Joined: 30 Nov 2004 Posts: 10315 Location: Córdoba (Spain)
|
Posted: Sat Oct 15, 2005 2:29 pm Post subject: |
|
|
@Superstoned, the topic is about if reiser4 should be included into gentoo-sources, not about if new tendencies sucks or not. I agree that there is a need for people to test such things, I tested it some months ago and you can be sure that, if I have a spare machine, that one is for testing. You would be surprised how many unstable software do I run. (No, I dont like gnome <1.4 and such ).
While all the filesystems fail, you have to understand that, under controlled circunstances, and assuming that the hardware is not going to fails, it is almost impossible to break ext3/2 (xfs is another totally different story), while the failures in reiser4 are more common, and most times, corruption is not due to a hardware failure, but a fs bug.
What we are discussing here is if reiser4 is mature enough to get into a stable branch of the kernel, Linus Torvalds says "no", and I agree.
We are on the same side, I say also let people try. But dont put that piece of code into a stable kernel unless it is mature and finished. And, please, dont use the reiser3, xfs, ntfs and whatever argument. We are here to learn from the past errors, not to repeat them. |
|
Back to top |
|
|
codergeek42 Bodhisattva
Joined: 05 Apr 2004 Posts: 5142 Location: Anaheim, CA (USA)
|
Posted: Sat Oct 15, 2005 8:31 pm Post subject: |
|
|
superstoned wrote: | reiser4 might not be very stable, as it is new. but it's a fcking filesystem, why all the fuzz? let ppl try it, why not, choice is one of the nice things FOSS has to offer?!? | Exactly. You have the choice to patch your own kernel with Reiser4 support, should you want to. However, Linus and other kernel hackers still think it's not quite stable enough, and not coded so well, for inclusion into the mainline kernel. _________________ ~~ Peter: Programmer, Mathematician, STEM & Free Software Advocate, Enlightened Agent, Transhumanist, Fedora contributor
Who am I? :: EFF & FSF |
|
Back to top |
|
|
brakits n00b
Joined: 13 May 2002 Posts: 10 Location: San Francisco, CA
|
Posted: Sat Jan 07, 2006 7:12 am Post subject: can we expect resierfs 4 to end up in gentoo sources soon? |
|
|
is the kernel team looking at putting reiserfs 4 in the main gentoo kernel anytime soon?
thank
-s _________________ --
Email: "Stephen Harrell <stephen@teknikal.org>"
WWW: http://teknikal.org
Photography: http://delayed.org
Public GPG Key: http://pgp.stephenharrell.com |
|
Back to top |
|
|
mark_alec Bodhisattva
Joined: 11 Sep 2004 Posts: 6066 Location: Melbourne, Australia
|
Posted: Sat Jan 07, 2006 7:25 am Post subject: |
|
|
Merged above. _________________ www.gentoo.org.au || #gentoo-au |
|
Back to top |
|
|
Paapaa l33t
Joined: 14 Aug 2005 Posts: 955 Location: Finland
|
Posted: Sat Jan 07, 2006 10:32 am Post subject: |
|
|
Yet another file system speed comparsion including Ext2, Ext3, ReiserFS, Reiser4, XFS and JFS:
http://linuxgazette.net/122/TWDT.html#piszcz
Remember that the test was made using a relatively slow computer, Pentium III 500MHz. Conclusion from the site:
Quote: | "With the second round of filesystem benchmarks, I hope everyone is now satisfied with the benchmarks using the 2.6 kernel. What I gleam from these benchmarks is both EXT2 and EXT3 are now roughly the same speeds in the majority of the tests. It also appears the XFS has improved in the majority of the tests. ReiserFSv3 has slowed in many of the tests with ReiserFSv4 being the slowest in most of the tests. It is important to note that JFS has improved in some of the tests. Personally, I still choose XFS for filesystem performance and scalability." |
No, I don't need Reiser4 to be included in gentoo-sources.
Last edited by Paapaa on Sat Jan 07, 2006 12:12 pm; edited 1 time in total |
|
Back to top |
|
|
StringCheesian l33t
Joined: 21 Oct 2003 Posts: 887
|
Posted: Sat Jan 07, 2006 11:03 am Post subject: |
|
|
Paapaa wrote: | Yet another file system speed comparsion including Ext2, Ext3, ReiserFS, Reiser4, XFS and JFS:
http://linuxgazette.net/122/TWDT.html#piszcz
Remember that the test was made using a relatively slow computer, Pentium III 500MHz. |
That Reiser4 mostly lost is not just because it was on a 500Mhz, it's also (I've heard) because Hans has been making performance compromises to meet the kernel dev's requirements for merging. |
|
Back to top |
|
|
superstoned Guru
Joined: 17 Dec 2004 Posts: 432
|
Posted: Sat Jan 07, 2006 8:02 pm Post subject: |
|
|
StringCheesian wrote: | Paapaa wrote: | Yet another file system speed comparsion including Ext2, Ext3, ReiserFS, Reiser4, XFS and JFS:
http://linuxgazette.net/122/TWDT.html#piszcz
Remember that the test was made using a relatively slow computer, Pentium III 500MHz. |
That Reiser4 mostly lost is not just because it was on a 500Mhz, it's also (I've heard) because Hans has been making performance compromises to meet the kernel dev's requirements for merging. | well, then, i hope he can fix this later on - tough its important to get reiser4 in the kernel, performance has been the selling point for reiser4... |
|
Back to top |
|
|
lefsha Veteran
Joined: 30 Aug 2004 Posts: 1234 Location: Burgas, Bulgaria
|
Posted: Sun Apr 02, 2006 1:21 am Post subject: |
|
|
Thanks to you all guys!
More then 60% staying for including Resier4 into Gentoo-Kernel.
That is nice. Developers need to think about it seriosily even if some of them
personally don't need (like) this FS.
If Gentoo is all about choices, so give us the possibility to decide whether we need
it or not!
Thanks once more. _________________ Lefsha |
|
Back to top |
|
|
djm Arch/Herd Tester
Joined: 12 Apr 2004 Posts: 690 Location: Wadham College, Oxford
|
Posted: Sun Apr 02, 2006 10:55 am Post subject: |
|
|
No one's stopping you _________________ the forums.gentoo.org poster formally known as metal leper |
|
Back to top |
|
|
amne Bodhisattva
Joined: 17 Nov 2002 Posts: 6378 Location: Graz / EU
|
Posted: Thu Apr 06, 2006 8:59 am Post subject: |
|
|
Uhm, that poll doesn't even have an option that says "no" - "i don't care" is something different than "no, i don't want reiser4 in gentoo-sources because it is against the principle of the Gentoo-sources".
Gentoo-sources is only vanilla-sources plus Bugfixes plus the speakup patches. Nothing else, no features added. Hence reiser4 has no place in there as long it's not part of vanilla sources. _________________ Dinosaur week! (Ok, this thread is so last week) |
|
Back to top |
|
|
superstoned Guru
Joined: 17 Dec 2004 Posts: 432
|
Posted: Thu Apr 06, 2006 4:54 pm Post subject: |
|
|
amne wrote: | Uhm, that poll doesn't even have an option that says "no" - "i don't care" is something different than "no, i don't want reiser4 in gentoo-sources because it is against the principle of the Gentoo-sources".
Gentoo-sources is only vanilla-sources plus Bugfixes plus the speakup patches. Nothing else, no features added. Hence reiser4 has no place in there as long it's not part of vanilla sources. | yeah, tought that was clear already. i love the idea of R4 (tough it doesn't live up to the hype - yet) but gentoo-sources should be stable. if it had more than these bugfixes, i'd ask for inclusion of -ck... |
|
Back to top |
|
|
slycordinator Advocate
Joined: 31 Jan 2004 Posts: 3065 Location: Korea
|
Posted: Fri Apr 07, 2006 6:03 am Post subject: |
|
|
Voltago wrote: | Problem with any community is, the screaming fanboys tend to be the ones everyone notices, naturally. |
But in the case of reiserfs, Daniel Robbins was the "screaming fanboy." Hell, I remember when the install guide specifically said that it was reccomended to have your / file system be reiserfs. |
|
Back to top |
|
|
StringCheesian l33t
Joined: 21 Oct 2003 Posts: 887
|
Posted: Fri Apr 07, 2006 7:12 am Post subject: |
|
|
As for those for whom R4 doesn't work perfectly - how is the open source many-eyes effect supposed to really start working its magic when the software is hidden away to the alternative kernels/LiveCDs and not so conveniently available?
Making it available on the LiveCD and kernel with a zero support disclaimer would be enough... |
|
Back to top |
|
|
djm Arch/Herd Tester
Joined: 12 Apr 2004 Posts: 690 Location: Wadham College, Oxford
|
Posted: Fri Apr 07, 2006 10:06 am Post subject: |
|
|
StringCheesian wrote: | As for those for whom R4 doesn't work perfectly - how is the open source many-eyes effect supposed to really start working its magic when the software is hidden away to the alternative kernels/LiveCDs and not so conveniently available?
Making it available on the LiveCD and kernel with a zero support disclaimer would be enough... |
You're assuming that people would have any idea how to create good bug reports (or even ones less crappy than "b0rk!") for their filesystem. I don't imagine many would, and I don't imagine many would want to bother.
The people who have the know how and the inclination are more than capable of using reiser4 without it being in g-s _________________ the forums.gentoo.org poster formally known as metal leper |
|
Back to top |
|
|
StringCheesian l33t
Joined: 21 Oct 2003 Posts: 887
|
Posted: Fri Apr 07, 2006 10:25 am Post subject: |
|
|
djm wrote: | The people who have the know how and the inclination are more than capable of using reiser4 without it being in g-s |
But there's so much developement/bugfixing that's proportional to user popularity. Paid developers working for commercial distros - businesses respond to demand - for just one example. So you see greater usage even by people without the inclination or know how can do far more, indirectly, for getting more attention from good testers/devs.
How could the popularity of R4 among Gentoo users not effect how rapidly R4 improves? |
|
Back to top |
|
|
djm Arch/Herd Tester
Joined: 12 Apr 2004 Posts: 690 Location: Wadham College, Oxford
|
Posted: Fri Apr 07, 2006 11:11 am Post subject: |
|
|
StringCheesian wrote: | djm wrote: | The people who have the know how and the inclination are more than capable of using reiser4 without it being in g-s |
But there's so much developement/bugfixing that's proportional to user popularity. Paid developers working for commercial distros - businesses respond to demand - for just one example. So you see greater usage even by people without the inclination or know how can do far more, indirectly, for getting more attention from good testers/devs.
How could the popularity of R4 among Gentoo users not effect how rapidly R4 improves? |
As far as I'm concerned, you're living in a fantasy world. Lots of Gentoo users using it won't do anything other than increasing the bad reputation gentoo users have and possibly creating bugs which aren't obviously filesystem related, but end up being because of some weird bug, after someone's spent ages trying to figure it out.
Anyway, read what amne said. That's all there is too it, really. _________________ the forums.gentoo.org poster formally known as metal leper |
|
Back to top |
|
|
F.Ultra Apprentice
Joined: 17 Mar 2004 Posts: 169 Location: Sweden
|
Posted: Mon Apr 24, 2006 11:09 am Post subject: |
|
|
Quote: | Yet another file system speed comparsion including Ext2, Ext3, ReiserFS, Reiser4, XFS and JFS:
http://linuxgazette.net/122/TWDT.html#piszcz
Remember that the test was made using a relatively slow computer, Pentium III 500MHz. | I decided to run the tests performed on that page since there has been quite a debate regarding the hardware used by the linuxgazette test.
The tests where performed on a HP DL145G2 with 2 AMD Opteron 275@2.2GHz and 4GB of memory (that is 4 processor cores at 2.2GHz), and the storage device was a Coraid Etherdrive SR1520 populated with 15 SATA drives with a total size of 1.2TB (I made a 1.2TB partition for this test, the total disk size is otherwhise 4.2TB). The Etherdrive exported the drives as a Raid 5 array to the server, and the Etherdrive was directly connected to the server using a crosswired cat6 network cable.
All tests where performed with the 2.6.15-gentoo-r7 kernel in 64-bit mode. Reiser4 was the latest patch from namesys for 2.6.15. Since the AoE driver in 2.6.15 is very old (v16 when the latest from Coraid is v27) I had no chance to perform this test with the Etherdrive at it's peak performance (the never drives supports jumbo frames among other things).
In the table below I have entered the time in seconds and each number is postfixed with the "position" where 1 is best and 6 is worst. The Points column is the sum of the positions so lower is better here as well, and in this sum I have exluded the mount time since I do not really feel that that should be included in the performance count even though the long mount times for Reiser3 and 4 can be quite significant for some.
I never timed the filesystem creation time since that is completely irrelevant, but ext2 and ext3 took well over an hour, jfs and xfs took a few seconds, Reiser3 and Reiser4 took around 10 minutes. Something that I noticed though was that after the final test (021) there was a significant long time for "sync+unmount" for both JFS and XFS, telling me that these two do some heavy caching.
Code: |
Reiser4 Reiser3 JFS XFS EXT2 EXT3
001 9.335(2) 9.288(1) 9.649(3) 10.587(5) 10.435(4) 13.300(6)
002 0.014(4) 0.005(3) 0.004(2) 0.004(2) 0.003(1) 0.004(2)
003 0.471(4) 0.389(3) 1.397(5) 3.794(6) 0.044(1) 0.148(2)
004 7.698(2) 7.588(1) 7.834(3) 9.123(4) 16.591(6) 13.334(5)
005 0.207(5) 0.234(6) 0.096(1) 0.105(3) 0.097(2) 0.130(4)
006 28.397(6) 11.224(5) 1.332(3) 3.618(4) 0.357(1) 0.812(2)
007 0.097(2) 0.078(1) 0.103(3) 0.104(4) 0.288(5) 0.318(6)
008 0.215(6) 0.164(3) 0.165(4) 0.158(1) 0.161(2) 0.171(5)
009 14.195(1) 15.066(4) 16.902(5) 14.884(3) 14.510(2) 22.075(6)
010 55.700(3) 55.938(6) 55.903(4) 55.579(2) 55.406(1) 55.919(5)
011 1.494(4) 1.071(3) 3.026(5) 4.135(6) 0.202(1) 0.487(2)
012 1.060(3) 0.810(1) 1.069(5) 1.068(4) 1.012(2) 1.729(6)
013 46.858(6) 7.394(2) 27.708(4) 3.629(1) 25.824(3) 39.475(5)
014 44.567(6) 2.338(1) 27.570(4) 3.009(2) 25.065(3) 38.342(5)
020 2.547(2) 24.176(6) 4.134(4) 3.403(3) 1.501(1) 12.445(5)
021 0.775(5) 0.588(1) 0.740(3) 0.837(6) 0.731(2) 0.757(4)
mount 126.153(5) 134.314(6) 0.004(1) 0.023(2) 0.117(3) 0.132(4)
--------------------------------------------------------------------------------------
points 61 47 58 56 37 70
Filesystem sorted by points:
#1 EXT2 37
#2 Reiser3 47
#3 XFS 56
#4 JFS 58
#5 Reiser4 61
#6 EXT3 70
|
I think that it is very hard to make any good conclusions from this results any than that they are different from the linuxgazette numbers which probably is due to the extreme changes in hardware. One conclusion is however that Reiser3 and Resier4 could do with better mount times
Reiser4 wins in test 009
Reiser3 wins in tests 001, 004, 007, 012, 014, 021
JFS wins in test 005
XFS wins in tests 008, 013
EXT2 wins in tests 002, 003, 006, 010, 011, 020
EXT3 wins no tests
Although Reiser3 wins the most tests, I would say that XFS looks very promising if one looks at the time diffs in actual seconds (I have not considered EXT2 here since I really do not think that an unjournaled filesystem is suited for larger partitions).
I had no possibility to perform tests 015, 016, 017, 018 and 019.
If someone would have the ability to make some graphs from this data and present in this thread it would be very much appreciated!
Last edited by F.Ultra on Wed May 03, 2006 9:04 pm; edited 2 times in total |
|
Back to top |
|
|
|
|
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
|
|