Prev: Compiz
Next: Confused about a software update
From: Hans on 17 Feb 2010 00:01 What are the pros and consadviseadvise (stability, error recovery, speed) of Reiser & ext4 on a Suse 11.2 system for boot, root, mail (mail dir), home partition (1,000's video files), advise please.
From: noi ance on 17 Feb 2010 13:45 On Wed, 17 Feb 2010 15:01:46 +1000, Hans typed this message: > What are the pros and consadviseadvise (stability, error recovery, > speed) of Reiser & ext4 on a Suse 11.2 system for boot, root, mail (mail > dir), home partition (1,000's video files), > > advise please. Well, ext4 is the default filesystem for Suse 11.2. Its supposed to be actually journal transactions that ext3 doesn't and do it faster than ext3. Reiser is definitely slower than ext3 and ext4. IIRC, Reiser was for very large filesystems and securer than ext3 but that is now covered by ext4 and faster. Recommendation is to go with ext4 on Suse 11.2, After the initial fschk ext4 systems will fschk much faster. There's a few routines to run to convert existing ext2 and ext3 filesystems to ext4 though not recommended for the /boot partition. Haven't found the routine to convert Reiser to Ext4 yet.
From: David Bolt on 17 Feb 2010 16:45 On Wednesday 17 Feb 2010 18:45, while playing with a tin of spray paint, noi ance painted this mural: > On Wed, 17 Feb 2010 15:01:46 +1000, Hans typed this message: > >> What are the pros and consadviseadvise (stability, error recovery, >> speed) of Reiser & ext4 on a Suse 11.2 system for boot, root, mail (mail >> dir), home partition (1,000's video files), >> >> advise please. > > Well, ext4 is the default filesystem for Suse 11.2. Its supposed to be > actually journal transactions that ext3 doesn't and do it faster than > ext3. And it is supposed to be more intelligent about allocating blocks so there's less chance of fragmentation, it supports online defragmenting, and some other features that most (all?) other file systems don't have. > Reiser is definitely slower than ext3 and ext4. I found that there wasn't much noticeable difference between ext3 and reiserfs in use. The ext3 file system was faster to mount but, when the number of files inside directories increased past several thousand, reiserfs became noticeably quicker at getting the directory listings. > IIRC, Reiser was for > very large filesystems and securer than ext3 but that is now covered by > ext4 and faster. If I was planning to deploy a new system, I'd use reiserfs for any file systems with lots of little files. For file systems with multi-GB files, e.g. working with raw video files or very large databases, I'd use xfs[0]. I normally use ext2 for /boot as have it as a separate system and there's no need for it to waste space with a journal. Finally, I'd use ext3 for general usage file systems. > Recommendation is to go with ext4 on Suse 11.2, After the initial fschk > ext4 systems will fschk much faster. There's a few routines to run to > convert existing ext2 and ext3 filesystems to ext4 though not recommended > for the /boot partition. Well, since there's no need for a journal on /boot, since it's very rarely going to be written to, using any file system with a journal just wastes space on it. > Haven't found the routine to convert Reiser to > Ext4 yet. You won't find a routine to do it other than to copy the contents from the reiserfs file system, use mkfs.ext4 to create the ext4 file system, and then copying the files back. The same applies when converting from jfs, xfs or any of the Windows file systems. [0] you may use jfs instead but, since I've never used jfs or looked into its strengths and weaknesses, I don't know whether it would be more suitable. Regards, David Bolt -- Team Acorn: www.distributed.net OGR-NG @ ~100Mnodes RC5-72 @ ~1Mkeys/s openSUSE 11.0 32b | | | openSUSE 11.3M1 32b openSUSE 11.0 64b | openSUSE 11.1 64b | openSUSE 11.2 64b | TOS 4.02 | openSUSE 11.1 PPC | RISC OS 4.02 | RISC OS 3.11
From: Hans on 18 Feb 2010 08:53 David Bolt wrote: > On Wednesday 17 Feb 2010 18:45, while playing with a tin of spray paint, > noi ance painted this mural: > >> On Wed, 17 Feb 2010 15:01:46 +1000, Hans typed this message: >> >>> What are the pros and consadviseadvise (stability, error recovery, >>> speed) of Reiser & ext4 on a Suse 11.2 system for boot, root, mail (mail >>> dir), home partition (1,000's video files), >>> >>> advise please. >> Well, ext4 is the default filesystem for Suse 11.2. Its supposed to be >> actually journal transactions that ext3 doesn't and do it faster than >> ext3. > > And it is supposed to be more intelligent about allocating blocks so > there's less chance of fragmentation, it supports online defragmenting, > and some other features that most (all?) other file systems don't have. > >> Reiser is definitely slower than ext3 and ext4. > > I found that there wasn't much noticeable difference between ext3 and > reiserfs in use. The ext3 file system was faster to mount but, when the > number of files inside directories increased past several thousand, > reiserfs became noticeably quicker at getting the directory listings. > >> IIRC, Reiser was for >> very large filesystems and securer than ext3 but that is now covered by >> ext4 and faster. > > If I was planning to deploy a new system, I'd use reiserfs for any file > systems with lots of little files. For file systems with multi-GB > files, e.g. working with raw video files or very large databases, I'd > use xfs[0]. I normally use ext2 for /boot as have it as a separate > system and there's no need for it to waste space with a journal. > Finally, I'd use ext3 for general usage file systems. > >> Recommendation is to go with ext4 on Suse 11.2, After the initial fschk >> ext4 systems will fschk much faster. There's a few routines to run to >> convert existing ext2 and ext3 filesystems to ext4 though not recommended >> for the /boot partition. > > Well, since there's no need for a journal on /boot, since it's very > rarely going to be written to, using any file system with a journal > just wastes space on it. > >> Haven't found the routine to convert Reiser to >> Ext4 yet. > > You won't find a routine to do it other than to copy the contents from > the reiserfs file system, use mkfs.ext4 to create the ext4 file system, > and then copying the files back. The same applies when converting from > jfs, xfs or any of the Windows file systems. > > > [0] you may use jfs instead but, since I've never used jfs or looked > into its strengths and weaknesses, I don't know whether it would be > more suitable. > > Regards, > David Bolt > Thank for the advise. I will go with EXT4 and change to Reiser for the Mail Dir partition if it gets to slow. Conversion is easy. I copy the directories and files to a spare (Backup) partition, format the original partition to the new file system, then copy the contents of the spare partition to the original partition with the new file system. regards from Down Under Hans
From: Chris Cox on 18 Feb 2010 12:26
On Wed, 2010-02-17 at 21:45 +0000, David Bolt wrote: > On Wednesday 17 Feb 2010 18:45, while playing with a tin of spray paint, > noi ance painted this mural: > > > On Wed, 17 Feb 2010 15:01:46 +1000, Hans typed this message: > > > >> What are the pros and consadviseadvise (stability, error recovery, > >> speed) of Reiser & ext4 on a Suse 11.2 system for boot, root, mail (mail > >> dir), home partition (1,000's video files), > >> > >> advise please. > > > > Well, ext4 is the default filesystem for Suse 11.2. Its supposed to be > > actually journal transactions that ext3 doesn't and do it faster than > > ext3. > > And it is supposed to be more intelligent about allocating blocks so > there's less chance of fragmentation, it supports online defragmenting, > and some other features that most (all?) other file systems don't have. ext4 is doing some of the SAME things that have been in modern filesystems (e.g. reiserfs) for MANY years... so yes... the ugly fragmentation issue which is throughout ext2 and ext3 is now minimized.... yes... indeed it's almost reiserfs like (I've never seen a reiserfs go beyond about 9% frag even after 10 years or more, can't say the same about ext2/3). > > > Reiser is definitely slower than ext3 and ext4. Sigh.. shame the benchmarks do NOT show this. Where they do perform better, it's not by significant margins IMHO. But if we keep chanting it over and over it will make it true. > > I found that there wasn't much noticeable difference between ext3 and > reiserfs in use. The ext3 file system was faster to mount but, when the > number of files inside directories increased past several thousand, > reiserfs became noticeably quicker at getting the directory listings. Duh.... but ext3 is god... and we're not allowed to challenge the mighty Red Hat. But again, ext4 fixes some of these issues.. in fact I think a lot of the later patches to ext3 fixed a lot of the lookup performance issues. > > > IIRC, Reiser was for > > very large filesystems and securer than ext3 but that is now covered by > > ext4 and faster. > > If I was planning to deploy a new system, I'd use reiserfs for any file > systems with lots of little files. For file systems with multi-GB > files, e.g. working with raw video files or very large databases, I'd > use xfs[0]. I normally use ext2 for /boot as have it as a separate > system and there's no need for it to waste space with a journal. > Finally, I'd use ext3 for general usage file systems. I agree with the comments on xfs. > > > Recommendation is to go with ext4 on Suse 11.2, After the initial fschk > > ext4 systems will fschk much faster. There's a few routines to run to > > convert existing ext2 and ext3 filesystems to ext4 though not recommended > > for the /boot partition. > > Well, since there's no need for a journal on /boot, since it's very > rarely going to be written to, using any file system with a journal > just wastes space on it. I use ext2 for /boot areas for this very reason. Journals require space (if not separated). > > > Haven't found the routine to convert Reiser to > > Ext4 yet. > > You won't find a routine to do it other than to copy the contents from > the reiserfs file system, use mkfs.ext4 to create the ext4 file system, > and then copying the files back. The same applies when converting from > jfs, xfs or any of the Windows file systems. > > > [0] you may use jfs instead but, since I've never used jfs or looked > into its strengths and weaknesses, I don't know whether it would be > more suitable. JFS is buggy. I would NOT recommend it. IBM could always open source their GOOD filesystem... but they won't :-) The main reason why ext* is pushed is because of the lack of development on reiserfs and the arrest of the primary "brain" behind reiser4. Is ext4 better than ext3? Sure... once all of the bugs are squashed out. Is ext4 better than reiserfs (a very old filesystem)? Eh... not really that much better at all. It's like somebody decided to try to do some catch up tech wise, but didn't want to surpass anything that was out there (except their own past versions). The futures however DOES NOT belong to ext4 (it's still pretty basic, not as feature full as reiserfs and certainly nothing like reiser4). Future filesystems that show promise are brtfs and nilfs2... and there are couple of others.... I suppose even reiser4 (if the developers that remain are motivated to continue work). |