Hi! I was just wondering how Ext3 and Reiserfs compare. When I reinstalled my server (because of a stupid hacker) I took the opportunity to change to ReiserFS. And I have to say it's really much faster than Ext3. I don't have benchmarks, but for example, stuff like "make dep" on the linux kernel is much faster (even though I had enabled write cache when I was using ext3). So what's some highlights on Ext3 vs. ReiserFS? I guess the Ext2 compability is one large factor for using Ext3, but otherwise? _____________________________________________________ | Martin Eriksson <nitrax@giron.wox.org> | MSc CSE student, department of Computing Science | Umeå University, Sweden
I am using ext3 on my production systems. My interests lie more in large file performance that small file performance. In this situation Reiser did not perform well. I was seeing 20-30 MB/s for a ResierFs system, but ext3 was giving me 70 MB/s for reads and 100 MB/s for writes. All reports I have read say that Reiser is very good at small files. You see this during your kernel makes. Craig On Tue, Jan 15, 2002 at 05:26:55PM +0100, Martin Eriksson wrote:> Hi! > > I was just wondering how Ext3 and Reiserfs compare. When I reinstalled my > server (because of a stupid hacker) I took the opportunity to change to > ReiserFS. And I have to say it's really much faster than Ext3. > > I don't have benchmarks, but for example, stuff like "make dep" on the linux > kernel is much faster (even though I had enabled write cache when I was > using ext3). > > So what's some highlights on Ext3 vs. ReiserFS? I guess the Ext2 compability > is one large factor for using Ext3, but otherwise? > > _____________________________________________________ > | Martin Eriksson <nitrax@giron.wox.org> > | MSc CSE student, department of Computing Science > | Umeå University, Sweden > > > > > _______________________________________________ > Ext3-users mailing list > Ext3-users@redhat.com > https://listman.redhat.com/mailman/listinfo/ext3-users-- Craig Tierney (ctierney@hpti.com)
#include <hallo.h> Martin Eriksson wrote on Tue Jan 15, 2002 um 05:26:55PM:> I was just wondering how Ext3 and Reiserfs compare. When I reinstalled my > server (because of a stupid hacker) I took the opportunity to change to > ReiserFS. And I have to say it's really much faster than Ext3. > > I don't have benchmarks, but for example, stuff like "make dep" on the linux > kernel is much faster (even though I had enabled write cache when I was > using ext3).I have to second this. My last impression of ReiserFS was: - needs less IO traffic for meta handling - means less! visible load on IDE systems - faster recursive searches - saves space. In our tests with /usr directories of typical systems, ReiserFS wasted <2% space. XFS wasted ~7% and Ext2/3 wasted 16.5 percent! (typical, with 4kB blocks) - Ext2/3 behave worse, Ext3 sometimes better, sometimes even more worse. The CPU load while massive disk write operations even sometimes makes my mouse and the window manager (which makes some IO handling periodicaly, I guess) freeze for few seconds. This did never happen with ReiserFS.> So what's some highlights on Ext3 vs. ReiserFS? I guess the Ext2 compability > is one large factor for using Ext3, but otherwise?- Stability. ReiserFS changed formats many times, often in incompatible ways. - Safeness. Ext3 should loose less data on crashes. - Same on hardware damages. In past, I had problems with different harddisks. e2fsck -c was very often successfully (means, could save what was not damaged), only if the directory node has been damaged, the file names sometimes got lost and I had to rename them. reiserfsck sucked at this point. Note there is also less reliable hardware - AFAIK on such systems Reiserfs sometimes is going broken. The situation may have changed with the recent ReiserFS code, but I would not trust them. My summary: use Ext3 for important data: /var, /home, /etc, / and ReiserFS for /usr and other parts which you can easily restore when something is broken. Gruss/Regards, Eduard. -- <Angel`Eye> installations anleitung für intelx86 richtig ? <Salz> Angel`Eye: Kommt auf deinen Rechner an. Wenn du die Antwort nicht weiß, ist sie ja. -- #debian.de
Craig Tierney wrote:> > I am using ext3 on my production systems. My interests lie more in > large file performance that small file performance. In this situation > Reiser did not perform well. I was seeing 20-30 MB/s for a ResierFs system, > but ext3 was giving me 70 MB/s for reads and 100 MB/s for writes. > > All reports I have read say that Reiser is very good at small files. You see > this during your kernel makes.try to look this: http://www.uwsg.indiana.edu/hypermail/linux/kernel/0112.0/1241.html regards,
On Tue, Jan 15, 2002 at 11:44:57PM +0100, Eduard Bloch wrote:> My summary: use Ext3 for important data: /var, /home, /etc, / and > ReiserFS for /usr and other parts which you can easily restore when > something is broken./usr is mostly reads. Does ReiserFS score more in that over ext3? If not, then why lose the compatibilty of ext3 as well as ease of switch-over ( from and to ext2, in case of problems )? regards, jaju
Craig Tierney wrote:> > I am using ext3 on my production systems. My interests lie more in > large file performance that small file performance. In this situation > Reiser did not perform well. I was seeing 20-30 MB/s for a ResierFs system, > but ext3 was giving me 70 MB/s for reads and 100 MB/s for writes. > > All reports I have read say that Reiser is very good at small files. You see > this during your kernel makes.Hi, You can find benchs reiserfs vs. ext3 on http://www.namesys.com/benchmarks/benchmark-results.html I did also similar benchs: http://www.linux-france.org/article/sys/ext3fs/Benchmarks/benchmark-mongo-2.txt and http://www.linux-france.org/article/sys/ext3fs/Benchmarks/index.html (this one is in French) Best regards, Liu
On Tue, Jan 15, 2002 at 05:26:55PM +0100, Martin Eriksson wrote:> I was just wondering how Ext3 and Reiserfs compare. When I reinstalled my > server (because of a stupid hacker) I took the opportunity to change to > ReiserFS. And I have to say it's really much faster than Ext3. > > So what's some highlights on Ext3 vs. ReiserFS? I guess the Ext2 compability > is one large factor for using Ext3, but otherwise?I ran some benchmarks recently to test performance of several Linux filesystems under heavey synchronous load (ie a mail server). For this, ext3 data=journal was nearly twice as fast as ReiserFS. See http://bruce-guenter.dyndns.org/benchmarking/ -- Bruce Guenter <bruceg@em.ca> http://em.ca/~bruceg/ http://untroubled.org/ OpenPGP key: 699980E8 / D0B7 C8DD 365D A395 29DA 2E2A E96F B2DC 6999 80E8
Zoiah writes:> On Tue, Jan 15, 2002 at 05:26:55PM +0100, Martin Eriksson wrote: >> I was just wondering how Ext3 and Reiserfs compare. When I reinstalled my >> server (because of a stupid hacker) I took the opportunity to change to >> ReiserFS. And I have to say it's really much faster than Ext3. >> =20 >> So what's some highlights on Ext3 vs. ReiserFS? I guess the Ext2 >> compabil> ity >> is one large factor for using Ext3, but otherwise? > > I ran some benchmarks recently to test performance of several Linux > filesystems under heavey synchronous load (ie a mail server). For this, > ext3 data=3Djournal was nearly twice as fast as ReiserFS. > > See http://bruce-guenter.dyndns.org/benchmarking/I see in your benchmark that EXT3 is actually performing better than EXT2. How is that possible? Because as far as I know EXT3 is just EXT2 + journalling which means more work for the HD. Looking at the benchmark on http://www.mandrakeforum.com/print.php?sid=1212&lang=en I see that EXT3 performs equal or slower than EXT2. Just wondering, Erik Smit
Hi, On Fri, Jan 18, 2002 at 01:23:11PM +0000, Zoiah wrote:> I see in your benchmark that EXT3 is actually performing better than EXT2. > How is that possible? Because as far as I know EXT3 is just EXT2 + > journalling which means more work for the HD.It depends very much on the workload. ext3 can often avoid seeks that ext2 has to do, because it can flush data out sequentially to the journal rather than having to seek to all the bitmap and inode blocks when writing out a change to disk. This is especially noticeable with some synchronised-IO benchmarks, where ext2 has to seek all over the disk for every IO request, whereas ext3 can just append a bit more to the journal. Cheers, Stephen
On Fri, Jan 18, 2002 at 02:07:39PM +0000, Stephen C. Tweedie wrote:> It depends very much on the workload. ext3 can often avoid seeks that > ext2 has to do, because it can flush data out sequentially to the > journal rather than having to seek to all the bitmap and inode blocks > when writing out a change to disk. This is especially noticeable with > some synchronised-IO benchmarks, where ext2 has to seek all over the > disk for every IO request, whereas ext3 can just append a bit more to > the journal.It is especially more try in the case of a mail spool where the individual files are frequently deleted before the writeback interval would cause them to actually get written to their real location on disk. -- Bruce Guenter <bruceg@em.ca> http://em.ca/~bruceg/ http://untroubled.org/ OpenPGP key: 699980E8 / D0B7 C8DD 365D A395 29DA 2E2A E96F B2DC 6999 80E8