Which Partition Types Are Superior? 283
digitalmonkey2k1 writes: "I am currently planning on running an Apache web server and a small ftp on my pc. There are so many file systems that Linux can support now that I'm not certain what ones should be used for certain features. If anyone knows of a comparison list between them, somthing to give a pro/con method of deciding the best sort of configuration It would be greatly appreciated."
What's wrong with ext2? (Score:1)
Just DON'T even think about vfat or NTFS!
What's your problem with NTFS? (Score:1)
Just DON'T even think about vfat or NTFS!
Perhaps you meant using an NTFS fs under Linux, however if not: What's your problem with NTFS? NTFS 4 and NTFS 5 are very feature laden, impressive filesystems, so I'm curious what your problem with them is, apart from perhaps that they're from Microsoft.
Re:What's your problem with NTFS? (Score:1)
Re:What's your problem with NTFS? (Score:1, Informative)
Re:What's your problem with NTFS? (Score:2)
You left unanswered the qusestion of what FAT/NTFS have to do with Registry rot.
Re:What's your problem with NTFS? (Score:1)
However, it's also been my experience that under NTFS is damn unstable under WinNT as well.
Case in point... A production system I was running several years ago crashed. (NT4 SBS, PDC) The system HD was fried but the data drive was OK (backed up on RAID anyway.) When I attempted to reinstall the OS, for some strange reason NT told me the permissions on the data drive made it unusable! I ended up having to devise a particularly dangerous registry hack (involving NT's rescue boot-disk and reinstallation procedure) just to get access to a filesystem that should have just been readable right off the bat. Needless to say the system was down for several hours.
In another job, as a contract sysadmin for anout forty different clients, I was dealing with *constant* fs problems with NT.
Really, I only mentioned vfat and NTFS because I've met many linux newbies who are nervous that if they don't use a windows compatible FS they will be locked into Linux forever. (not that that's such a bad thing)
While ACL's are great, you can certainly use several different patches on ext2 and ext3 for those, or even use XFS. For a small web server, though, I don't think he needs anything more than ext2.
Just my $.02.
Partition != File System (Score:5, Insightful)
Besides, wasn't there just a story on Linux Advanced File Systems [slashdot.org]????
There is no reason why use shouldn't use ReiserFS. Performs just as well (for the most part) as ext2 and will improve uptime by eliminating nasty fscks in the event of a failure.
The more interesting question probably is what kernel should you be using????
Kudos to story posters for having so many dups today. Really keeping us on our toes huh?
Re:Partition != File System (Score:2, Interesting)
And no, tar is not an option, no matter what some Nameless Kernel Guru says. I've found that tar and cpio are awfully fragile and resource hoggish when it comes to backing up lots of data.
Re:Partition != File System (Score:1)
Re:Partition != File System (Score:2)
This is insightful? I kinda figured a post had to be RIGHT to be insightful...
There are several reasons not to use Reiser. There are lots of reasons to use it, as well, but you can't claim there are no reasons not to use it.
One reason not to use Reiser might be that you're building a box that doesn't need it. A firewall with a very small hard drive and a good power backup, for instance.
Another might be vendor support for your specific distribution. RedHat supports ext3 better currently, for instance.
Just because their FAQ says there are no reasons doesn't mean there aren't any.
Re:Rescue Floppy Support for ReiserFS? (Score:2)
VMWare is pretty cheap and plex86 is free (as is Bochs) if you need to run stability-compromising software as root (create a second account otherwise). Now you can have stability *and* a modern filesystem.
This is becoming a FAQ (Score:2, Offtopic)
Re:This is becoming a FAQ (Score:1)
/s
Moderators: why is this rated "insightful"?
Short answer: (Score:4, Insightful)
If you're looking to set up a small Apache and FTP for your friends, the short answer is probably: WHO CARES.
Ext2 will be more than sufficient, and if you feel like it just download the latest Redhat (7.2) and get Ext3 default for free. It probably won't make much difference to you for this task unless you're in an area powered by thousands of gerbils on wheels who happen to get tired at the same time.
Re:Short answer: (Score:1)
I do not think that Afghanistan has internet access at this time... ;P
Re:Short answer: (Score:2)
Come to think of it, maybe it was a look at Slashdot that lead them to ban the entire internet.
Re:Short answer: (Score:2, Informative)
I converted my slackware 8.0 system to ext3 in about 30 minutes. It's as simple as compiling a kernel with ext3 support, run tune2fs -j -Jsize=10
Re:Short answer: (Score:2, Insightful)
If you install a vanilla kernel, the machine will not be able to mount the root filesystem when you reboot (since it doesn't know anything about ext3 filesystems).
ext3 filesystems can be mounted by ext2 kernels as long as the fs was cleanly unmounted. (Doesn't help you after a crash, though, there you're right) That's the nice thing aout ext3. That and conversion to ext3 on the fly. And that it claims to be as fast or faster than ext2
Re:Short answer: (Score:3, Informative)
I don't really even know why someone would want to use ext3 anyway. Unless they've made some serious improvements in the past few months, the filesystem still writes at 50% the speed of ext2 (since all data is written twice). The only thing it has going for it is its interoperability with ext2, but that's really a perfidious "feature": systems that don't need journaling should just use ext2 to avoid the massive performance hit, and systems that do need journaling (namely, servers) have no reason to have their journaling filesystems compatible with ext2, and should use one of the high-performance journaling FS's.
Speed improvements in ext3 (Score:2)
They have made improvements, and in my experience ext3 is faster than ext2. See for example the Michael Johnson [linuxtoday.com] email:
Re:Short answer: (Score:2)
Not everyone realizes that the changes made to ext2 to add the journal are backward compatible. Don't be a dick.
Also, the point he makes about the root filesystem may not be 100% accurate, but the problem *will* arise if you have more than one partition. That is, you might be able to get / to mount when you reboot, but when the init scripts try to mount
the filesystem still writes at 50% the speed of ext2 (since all data is written twice).
Not true. Only one of the journaling modes does this, and it's not the default. The default mode only writes fs data twice, and is usually much faster, since it optimizes drive head motion by ordering the writes.
- It is better to remain silent and be thought a fool, than to open your mouth and remove all doubt.
Re:Short answer: (Score:4, Informative)
You're wrong. You didn't read what I said. (Score:3, Informative)
When Ext3 was first created, it COULD NOT journal metadata -- the only option was full file journaling, which was incredibly slow. Don't tell me I'm wrong, because I read the original release notes which said that metadata journaling was not available yet. I believe that Ext3 can now do metadata-only journaling -- somebody correct me if I'm wrong -- but it's a fairly recent development, within the past year or so.
This mail message [lwn.net] from about a year ago says that metadata support was "in an early state" at the time. I don't know if it's been perfected since then or not. But the e-mail proves that at one time, Ext3 could NOT do metadata-only journaling, which flat-out disproves your post that all journaling filesystems only journal metadata.
And RAID quite frankly has nothing to do with it; I can't even imagine why you brought it up because it's absolutely irrelevant.
Get a journaled FS (Score:5, Interesting)
XFS and ReiserFS are the more mature fs IMHO (on Linux) . I run EXT3 on systems that were previously running EXT2, because it's easy to upgrade. But I had some troubles with EXT3 not so long ago (corrupted files during a compilation, not even after a crash) .
ReiserFS is the best if you have a lot of small files. Both for performance and space. XFS is believed to be better for large files.
Also, if you need performance, FS is one thing, but software is another thing. Apache is probably the slowest web server out there (although very powerful (altough less than Roxen and Caudium
Running Zeus, Tux or (for static content) WebFS will give you a huge performance increase, even on a slow filesystem.
Re:Get a journaled FS (Score:2)
It's plenty fast enough if you connection to your web "customers" is through a T-3 or smaller - it will pump out data, even from a mod_perl script, fast enough to flood the link. I've never put it on a faster link so I don't know how much faster a link it will flood.
Re:Get a journaled FS (Score:1)
/s
Re:Get a journaled FS (Score:2)
Why not journal? What could you possibly lose by running a filesystem that protects the integrity of your data better, and runs faster?
As an aside, if he's exclusively serving data out, then the filesystem should be explicitly mounted read-only. Security and safety all at once!
Re:Get a journaled FS (Score:2, Informative)
It is for this reason that I'm switching my machines to ext3. IMHO corrupted files after a crash are just as intolerable as a corrupted filesystem. (ext3 does have a reiserfs-like metadata-only journaling mode, but by default it journals everything - at a small performance cost of course).
Re:Get a journaled FS (Score:2)
Re:Get a journaled FS (Score:2)
But journaling data doesn't get you too far if the programs you use aren't clever enough. Far too often an editor replaces file during saving instead of first writing a new one, after that removing old one and renaming new file to old filename. If your editor doesn't do this you're pretty much fscked up whether or not your filesystem supports journaling of data or not.
On the other hand if you're editing big file the feature that editor replaces the file is must because you might not have enough disk space or quota to store it twice. The same applies for databases and stuff that needs to keep files open for acceptable performance. If an editor does write changes to new file first and only after that removes old file and still data is missing after crash it's simply unacceptable. There should be very few cases where data journaling actually helps.
Re:Get a journaled FS (Score:2)
No, this is the correct way to update files. If you follow your instructions, then the permissions, any hard links will be lost.
you don't get something for nothing (Score:3, Informative)
Other problems with journaling file systems are that they are more complex, less mature, and have appeared only more recently in the Linux kernel, meaning there is a higher probability that they have some problem.
If you can't tolerate the few minutes of downtime resulting from an fsck, then a journaling file system is not going to help you either since machines become unavailable for lots of other reasons. In that case, you need network mirroring with a hot failover. Journaling file systems are more about convenience than any particularly rational engineering tradeoff.
Altogether, my recommendation is: don't pick software just because it's hot and new. For most users, ext2 with Apache makes a great web server platform. Apache is fast enough for any kind of Internet connection you are likely to have (Microsoft could probably serve all their static content from a single Apache server). If you like the convenience of a journaling file system and don't mind the performance hit, maybe you want to consider ReiserFS, which offers a lot of other useful features.
Re:you don't get something for nothing (Score:2)
You can waste a very large amount of time checking a filesystem after the system goes down unexpectedly (for whatever reason).
A journaling filesystem adds very little overhead, except when a large number of files are created, or deleted. As such, it is NOT going to have a large impact to a well configured webserver
Re:you don't get something for nothing (Score:2)
Re:you don't get something for nothing (Score:2)
I wouldn't use ext2 for any server even if I didn't need journaling. ReiserFS should be faster during reading especially if you happen to have many small files or a couple huge files. Journaling shouldn't matter performance if you aren't updating anything. And if you're, you should be using journaling anyway. So in the end, if you choose ReiserFS over ext2 you should get at least better performance. You might also get journaling as an added bonus.
I don't know if ReiserFS is the best journaling filesystem, but it's the one I have used and it has never disappointed me. OTOH, I have lost ext2 partition once... The only reason to use ext2 is either you're really low on memory (kernel size) or you have to use ext2 for backwards compatibility - for example I know about driver to read ext2 partitions from windows but I haven't found a one for reiserfs. Another way is obviously a piece of cake.
Re:Get a journaled FS (Score:3, Informative)
Re:Get a journaled FS (Score:2)
However, productions servers are usually not updated every day (especially the kernel), so XFS with a working kernel is ok.
But for workstations, ReiserFS may be a better choice, as it's in the kernel.
Re:Get a journaled FS (Score:2)
Which I would choose (Score:2, Interesting)
1. Data protection - the journaling systems Reiserfs, Ext3, XFS; etc offer far better and faster recovery than Ext2 does
2. Configurable, though I've never found a need to do it, you can set the block sizes in Ext2 and 3 to optimize it for larger files or smaller ones
What to stay away from fat/vfat/fat32. Why?
1. No security, period
*clink, clink* just my two cents..
Re:Which I would choose (Score:1)
This is less a performance enhancement and more a hack because of the finite amount of available inodes in ext filesystems.
ReiserFS dynamically allocates inodes so such as hack is unnecessary. ext filesystems are no more configurable than ReiserFS.
In fact, one can make very few arguments for using ext over ReiserFS. As reiser matures, I don't think there will be any reason to use ext2.
File system is half the disk story (Score:2, Insightful)
ReiserFS (Score:2, Interesting)
The best part is the journalling, if your machine with a 50 gig drive loses power and reboots, you do not have to go through a lengthy fsck, this would greatly decrease the boot time of your server.
Ext3 with standard dos-type partitions (Score:5, Informative)
Ext3 is basically ext2 with journalling. It performs better than Ext2, though. In a pinch you can always mount it as ext2.
You're not running anything exotic. Stick with the standards.
Re:Ext3 with standard dos-type partitions (Score:1)
DOS-type (6, e) partions will not support a linux filesystem.
I imagine your not referring to FAT32 (b, c) since that is in no way a DOS partition whereas FAT16 was just an extension of the original DOS filesystem much like the PE executable is just an extension of the MZ executable.
Re:Ext3 with standard dos-type partitions (Score:2)
In contrast, you can have BSD-style disk slices or some other partitioning scheme that bears no relation to what DOS uses.
Re:Ext3 with standard dos-type partitions (Score:2)
He was suggesting using the standard DOS partitioning scheme (you know -- up to 4 primary partitions, one of which can be turned into a logical/extended partition for more partitions) with an ext2 filesystem put onto these partitions.
There are other options -- you could just mke2fs /dev/hda if you wanted to, and you could use one of the other available partitioning schemes if you wanted -- but using the standard DOS scheme probably makes the most sense for most people.
Reiserfs, if you've got good hardware (Score:2, Informative)
As to reliability: if you've got good hardware, there shouldn't be any troubles at all. I for one, don't have good hardware.
Their repair tools suck, by the way.
So why do I keep using it? It's fast, is suse's default filesystem and it's fast.
NO, NO, NO! (Score:2, Insightful)
Which is the best car?
Who is the best actor?
Where is the best place to live?
None of these questions can be answered without saying "It depends" and neither can yours. Very rarely is anything better than everything else is every single way.
Re:NO, NO, NO! (Score:1)
Re:NO, NO, NO! (Score:2, Insightful)
Watermelon is better than every other fruit in every single way.
Re:NO, NO, NO! (Score:2)
"Obvious solution: genetically engineer a watermelon in a container similar to a banana. Result: the Waternana!
Or, depending upon what was meant by ease of use, the dildo-melon.
ReiserFS has distinct advantages. (Score:5, Informative)
I serve very few dynamic documents - I'm getting alot of milage out of small machines. My sites have a deep directory structure, with fairly few files in each. ReiserFS shines for this layout.
I tested several different FS for this application, ReiserFS won for me.
Oh yeah, the other benefit is the relative ease of install and upgrade.
journalized fs (Score:1)
I personally use Reiser FS on my "home surfing/programming" machine... because of the extensive support (especially in certain distros).
My preferred partition type is... (Score:5, Funny)
Of course, there are those who are type 82 bigots. I can see how that's important, but with RAM prices so low these days....
Matt
Re:My preferred partition type is... (Score:1)
Re:My preferred partition type is... (Score:1)
Re:My preferred partition type is... (Score:1)
Matt
Re:My preferred partition type is... (Score:1)
Partition type is different from partition format. ReiserFS, XFS, ext2, ext3, JFS are all partition formats and are the filesystems used on the disk. Partition type is only a number in the partition table, and in their Linux incarnations, all the aforementioned filesystems use type 83, simply indicating that partition is meant for Linux. You can call a partition type 69, and still format it with ext2 for all that it matters.
Re:My preferred partition type is... (Score:3, Interesting)
Of course the Amiga "RDB" partitioning scheme had its nice points too. Linux can read it, but I don't know if there are any Linux tools to create one.
ext stuff (Score:2)
My choices (Score:2, Interesting)
I did have a vfat drive (40 GB) that had a whole bunch of stuff from a when I used windows (98 SE then 2000 then XP) then I reinstalled and used it as a secondary drive. Worked for a few months until the partition table became corrupt, NOrton couldn't fix it and well here I am. (It is now a 40 GB reiser fs partition.)
Secondsun
Shouldn't be a difficult choice (Score:2, Funny)
Best of all, you can fully utilize it under Linux as well as Windows 2000, so if you feel like you would be better off with developing under Windows, you wouldn't have to reformat your whole disk and lose data in the process. Benchmarks have consistently shown [microsoft.com] that it is an enterprise-class performer.
Finally, you have to consider reliability in decisions such as these. NTFS just doesn't lose data, which is more than we can say of such "lossy" systems such as ReiserFS. Frankly, I can't even see why people put such "journaling" systems on production machines. All in all, you can't go wrong with NTFS.
Re:Shouldn't be a difficult choice (Score:1)
Re:Shouldn't be a difficult choice (Score:1)
#1 ntfs doesnt work under linux
#2 there are "free" systems that have ACLS (i believe XFS is one such)
#3 ext2 supports encryption i believe
and #4 NTFS is journaled i believe.
Re:Shouldn't be a difficult choice (Score:2)
Then why is the Linux write-drivers marked "dangerous" and "back up your NTFS volume first, because it will probably get damaged"[!!!]? Trying to run Linux off NTFS just sounds like a bad idea.
Whats wrong with google? (Score:1)
Re:Whats wrong with google? (Score:2, Informative)
ACLs on ext2, ext3, xfs (Score:5, Informative)
The XFS command line utilities seem to be less effective than the Bestbits patches & utils, and the Samba 2.2.1a support seems to be a bit off with its handling of recursive descents and inheritance. To be fair on both counts, I'm still learning the file system, and the problems could be all mine.
I'd thought about ReiserFS, but I really need those ACLs.
Just some thoughts. Any errors are all mine. Please feel free to correct. I have no pride.
ReiserFS is better. (Score:5, Informative)
Try ReiserFS. Too bad RedHat 7.2 decide not to support ReiserFS, I will give up - with regreat - on RedHat.
Re:ReiserFS is better. (Score:1)
2) Just because RedHat doesn't implement ReiserFS "natively" does NOT mean you cannot use it, nor does it mean that RH "decided" not to support it. They simply chose a different journaling FS which, imho, works pretty well.
Are there any other distributions that implement Reiser as default?
Re:ReiserFS is better. (Score:1)
I mention RedHat 7.2 because they seem to avoid European stuff many times, after all ext3 is the same as ext2. ReiserFS is much faster when you don't want file fragmention. ext3 was not designed from ground up and looking for file access time as goald one.
PC partition types suck (Score:1)
Whats the go with 4 primary partitions, the nextended partitions, some OS's primary partitions have to overlap extended partitions, some they done, some you can only use 1 primary partition, some you can use them all.
And having to reserve the whole first track, usually 32 sectors, just to put in about 170 bytes of data... what a waste.
The extra space can be used by other disk externder type programs, but there is no standard way to reserve it or anything, its first in best dressed, use it at your own risk.
And look at what it has to do, it just has to mark the start and end of a section of raw disk space, how difficult is it ?
The format has just been butchered too much by too many people. But for practical reasons we are stuck with PC partition tables because of convenience for the masses.
You could do a raw LVM partition.
I like Acorn partition types (if thats the one im thinking of), nice and simple, only what you need.
Let's not confuse FS and Partition type (Score:5, Informative)
IMHO, if you want a superior partition scheme, you should not use the linux system, which is identical in structure to the Microsoft DOS system. Instead, read about the BSD partition (and slice) system. See section 2.5.2 of the (new) 2d edition handbook:
http://www.freebsd.org/doc/en_US.ISO8859-1/books/
In BSD, the Microsoft-Linux concept of partitions is preserved as "slices" that exist to hold collections of files systems. (In FreeBSD, you can man hier(8) to read more about this. Unlike linux, where every vendor puts things in
Another option in BSD is the use of what are called "dangerously dedicated" system where the entire disk becomes one slides, with no other partition. Read more about this in the handbook. There's even information about working with different flavors of partition types.
I suppose to give 'equal time' we should give a link to the Microsoft/Linux partition scheme, so here one is:
http://www.redhat.com/docs/manuals/linux/RHL-7-Ma
FYI-- here's some news you won't see on
Re:Let's not confuse FS and Partition type (Score:1)
(e.g fields extracted from http://www.tac.eu.org/cgi-bin/man-cgi?disklabel+5
num of spare sectors per track
num of spare sectors per cylinder
num of alt. cylinders per unit
rotational speed
hardware sector interleave
sector 0 skew, per track
sector 0 skew, per cylinder
head switch time, usec
track-to-track seek, usec
cmon, do we need/want this information with modern storage systems, the details are taken care of in hardware, we dont need to clutter up our partition tables.
Not something to boast about.
lame slashdot lame filter prevented acurate formating
"dangerously dedicated" is that (Score:2)
It 'was' dangerous....if you did any kind of CVSup and rebuilt your kernel, things like top would stop working. SCSI drives formatted on DPT RAID cards would boot with an error on Adaptec cards, and a upgrade from 3.x to 4.x would break.
If you ran in a non-dedicated mode, rebuilds had no effect. Same with the SCSI issue.
Re:Let's not confuse FS and Partition type (Score:2)
Actually, the question asked about filesystems, quite clearly. The headline asked about partition types. Who comes up with the headline--the poster, or Hemos?
Re:Let's not confuse FS and Partition type (Score:5, Funny)
C) Cowboy Neal
Re:Let's not confuse FS and Partition type (Score:2)
Even though the question is obviously about filesystems...
The problem with BSDs partition is that most people run BSD on x86 boxes these days, so they end up putting it all inside of an old-fashioned "fdisk" style partition. Then you get the worst of both worlds. And if you go dedicated, everyone looks at you funny. "ARE YOU SURE YOU WANNA DO THAT?" Good grief, what are you supposed to say?
If you wanna see a partitioning system that really rocks, go back to the mid-80s and you'll find something that's better than anything else in use today: Amiga's RDB. And to think that this will fade into oblivion! Gee, thanks for reminding me how badly dilapidated and decayed personal computer technology is becoming. Here I am all happy that daylight savings time has finally ended and I get my stolen hour back, and someone has to remind me of a downer.
Using a PC these days is like touring Dunwich and Innsmouth.
Doesn't matter. (Score:5, Informative)
I'm running apache and ftp right now, and average traffic is about 20 hits per day. At this order of magnitude, or anywhere near it, it really doesn't matter.
Re:Doesn't matter. (Score:1)
And of course, the easiest way to change that is by telling people about it in Slashdot. Smart move. Good thing you didn't post the URL.
ext3 (Score:2, Insightful)
chattr +a
Of course, it has 1 big draw back, its not merged with Linus's kernel. That should be happening soon, I hope.
Personally I prefer (Score:1, Funny)
(heh, there is one in every crowd...and yes, that is me they are talking about)
Ok, go a head and mod this down as funny.
Another redundant question. (Score:1)
http://slashdot.org/article.pl?sid=01/10/23/162
appeared. And the "debate" over file systems has been going on and on. The article I cited above is recent and perhaps not as relevant, but I think a lot of people who wrote replies were both sensible and informative when it came to the merits of the two most available journaling systems for Linux (as is the case with this debate. I don't see a whole lot of chilish flaming.)
Some Important Tips to Remember (Score:4, Funny)
Partition? (Score:1)
My question is how then should we call those "parts" ?
Now to answer the filesystem question I would suggest using ReiserFS. I used it for a couple of months with good results and without any problem. It works well also on software raid1...
web server (Score:1)
it doesn't matter (Score:1, Redundant)
My fstab (Score:1)
/dev/hda3 / ext3
/dev/hda2 swap swap
/dev/hdb1
/dev/hda1/ / tmp reiserfs
The Default (Score:1)
My system would crash often (Score:1)
Which Partition Types are Superior (Score:1)
I personly like JFFS. Open source available from IBM. I also like JFFS in conjunction with LVM. LVM is awesome. Its snapshot feature is very useful in backing up active systems.
A crypto filesystem? (Score:4, Interesting)
I found a "CFS", but the version was just for kernel 2.2.x. I didn't find a 2.4 port.
SuSE 7.3 ads say it has a "CryptoFS". Does it work well? Where can I get it, if I don't want to install SuSE?
An easy-to-use crypto fs would be enormously important especially for laptops in corporate world. I think W2k or XP have some kind of encryption options, and if Linux can't provide a good alternative, it may be a problem in more paranoid companies.
Of normal filesystems, I've found ReiserFS stable on my two machines during my 6 months of use. I converted from ext2 after it corrupted mysteriously. Unfortunately, RH still doesn't support ReiserFS, even optionally, which I think is really silly. SuSE and Mandrake do.
Re:A crypto filesystem? (Score:2)
Re:A crypto filesystem? (Score:2)
If it's the CFS originally by Matt Blaze, then it should work fine without any kernel support other than NFS client. CFS just runs as a user-mode NFS server. If you can mount an NFS filesystem, you can mount a CFS file system. CFS even works on non-linux systems, such as FreeBSD.
My message of peace (Score:3, Funny)
Can't we all just thrash along?
--Blair
ReiserFS... (Score:3, Informative)
Example: You write a 513-byte file to a filesystem with 512-byte sectors. On other FS types, that file will take 2 sectors. On Reiser, it will take 1 sector plus change. Numerous small files of this type can have their tails packed into the same shared sector. I do not know the overhead in bytes per file, and thus don't know how many tails you can put into a given sector.
It also handles a very large number of files in the same directory well. Most other FS types have problems if you dump 10,000 files into a directory. It is my understanding that Reiser deals with this extremely well.
However, there is one drawback. If you are using LILO, the tail packing can cause you much grief. Lilo does not understand tails. It will be unable to execute its own second part or the kernel itself if either has had a tail-pack done. Thus, you should likely use a separate
Mandrake 8.0 came with a 2.2 kernel with ReiserFS backported. DO NOT try to use ReiserFS with any software RAID in any Linux 2.2 kernel. Make sure you update to 2.4. I believe 8.1 comes with 2.4 standard, so it shouldn't be an issue anymore with that distribution.
There have also been numerous bugfixes in the Reiser code over the 2.4 releases, so you will probably want to go with as recent a kernel as you can. Linus' 2.4 kernel tree has the reputation of being unstable, so you may want to use Alan Cox's branch until the official tree stabilizes better.
Options are just as important (Score:2)
Going mirrored will give the OS the opportunity to get the data from either spindle on reads.
just my $0.02
If you are talking about performance... (Score:2)
reiserfs, xfs and caching (Score:2, Interesting)
A simple solution is of course to disable swap, but that is only a workaround. Until reiserfs gets a little lighter on caching, my recommendation is definitely xfs.
Re:Upgrade to Dynamic Disk? (Score:2)
Dynamic Volumes == Software RAID. The data is stored on there, but the partition tables and FAT are so mangled that converting back and forth, or even just removing one of the drives and swapping in another, will render all data across all volumes unreadable.
MSKB has an article about this, I don't remember the number.
JKoebel