Re: ext3 in 2.4.14, or 2.4.15 source package?
Hi, Karsten. I'm Kazuhiko Uebayashi. I used kernel-patch(patch-2.4.15-pre3.gz from http://www.linuxhq.com/ ,it contains ext3 patch) ,using config-file kernel-2.4.13+ext3patch. It works well. On Fri, 16 Nov 2001 23:54:27 -0800 Karsten M. Self kmself@ix.netcom.com wrote: I'm looking for either an ext3fs patch to the 2.4.14 kernel sources (the 2.4.13 patch errors out), or would like the 2.4.15 kernel sources, which don't appear to be available yet -- any idea on when they're going to be out for Sid? Peace. -- Karsten M. Self kmself@ix.netcom.com http://kmself.home.netcom.com/ What part of Gestalt don't you understand? Home of the brave http://gestalt-system.sourceforge.net/ Land of the free Free Dmitry! Boycott Adobe! Repeal the DMCA! http://www.freesklyarov.org Geek for Hire http://kmself.home.netcom.com/resume.html -- Kazuhiok Uebyashi e-mail : [EMAIL PROTECTED]
Re: ext3 in 2.4.14, or 2.4.15 source package?
On Fri, 16 Nov 2001 23:54:27 -0800 Karsten M. Self kmself@ix.netcom.com wrote: I'm looking for either an ext3fs patch to the 2.4.14 kernel sources (the 2.4.13 patch errors out), or would like the 2.4.15 kernel sources, which don't appear to be available yet -- any idea on when they're going to be out for Sid? The ext3 patch was merged into 2.4.15pre2 (according to changelog). It's up to 2.4.15pre5 last I saw. So, you could try that, or wait for 2.4.15 which probably won't be too long in coming. As far as I can tell 2.4.15 is mostly a few smaller fixes and driver fixes/additions (apparently the loopback filesystem is fixed). 2.4.14 has been working pretty well for me so far ... YMMV. -- Eric G. Miller egm2@jps.net
Re: ext3 in 2.4.14, or 2.4.15 source package?
Karsten M. Self kmself@ix.netcom.com writes: I'm looking for either an ext3fs patch to the 2.4.14 kernel sources (the 2.4.13 patch errors out), or would like the 2.4.15 kernel sources, which don't appear to be available yet -- any idea on when they're going to be out for Sid? The 2.4.15 sources are still pre-release. I have yet to see a pre-release for 2.4 packaged in debian. Sid has kernel-patch-ext3-2.4, which replaces kernel-patch-2.4.9-ext3fs, kernel-patch-2.4.10-ext3fs, kernel-patch-2.4.13-ext3fs. Try that one out. -- Brian Nelson [EMAIL PROTECTED]
Re: ext3 in 2.4.14, or 2.4.15 source package?
on Sat, Nov 17, 2001 at 01:52:26AM -0800, Eric G. Miller (egm2@jps.net) wrote: On Fri, 16 Nov 2001 23:54:27 -0800 Karsten M. Self kmself@ix.netcom.com wrote: I'm looking for either an ext3fs patch to the 2.4.14 kernel sources (the 2.4.13 patch errors out), or would like the 2.4.15 kernel sources, which don't appear to be available yet -- any idea on when they're going to be out for Sid? The ext3 patch was merged into 2.4.15pre2 (according to changelog). It's up to 2.4.15pre5 last I saw. So, you could try that, or wait for 2.4.15 which probably won't be too long in coming. As far as I can tell 2.4.15 is mostly a few smaller fixes and driver fixes/additions (apparently the loopback filesystem is fixed). 2.4.14 has been working pretty well for me so far ... YMMV. I'll go this route. I just built 2.4.14 with reiserfs support, which is the bigger pain anyway (dancing the Reiserfs shuffle with large filesystems). Do that while the system isn't too populated, you can get stuff on and off partitions readily. The ext2 - ext3 switch is a lot easier to manage, I'll just have to build another kernel when 2.4.15's out. My current MO is to use ext3 for any FS ~256 MB. The 32 MB overhead of reiserfs just cuts into storage too much otherwise, and doesn't buy too much. Larger filesystems (which I want to spell feilsystems this hour of the night, it's another reiser bug...) get reisered. Seems to be a good balance. There's also a few new lightweight filesystems in 2.4, including a couple of ROM fs's I'd like to mess with. Peace. -- Karsten M. Self kmself@ix.netcom.com http://kmself.home.netcom.com/ What part of Gestalt don't you understand? Home of the brave http://gestalt-system.sourceforge.net/ Land of the free Free Dmitry! Boycott Adobe! Repeal the DMCA! http://www.freesklyarov.org Geek for Hire http://kmself.home.netcom.com/resume.html pgp7UahMQH43X.pgp Description: PGP signature
Re: So: reiserfs or ext3 (was Re: ext3 to be in 2.4.15!)
#include hallo.h Karsten M. Self wrote on Sat Nov 10, 2001 um 02:48:33PM: Now that you said this... I'd like to see how reiserfs and ext3 in writeback mode perform. I'd think ext3 would still be outperformed by reiser for the large dir listings. The issue isn't writing the directories, but searching them. Hash beats linear scan, above the hash overhead threshhold. This crossover's clearly evident for 10,000 directory entries, and is probably significant at substantially fewer. Please first how the thing works before adjucating relying on the traditional prejuducises (like speed: hash tree = list). This ist not allways true if you know how the client program work. In case of ext2, there were allways a problem: when doing file access in a directory (readdir), the ext2 driver did allways start at the top of the list and searched for that entry. Ext3 stores the position the last found entry, so when accessing again, the search begins an this position and you get your entry wery fast, in best case within the first search loop. Result: many traditional programs which read the directory contants in their natural order (and sort them later on displaying) work dimensions faster in large full directories than with the default ext2 driver. Btw, this modification could also be ported to ext2, I wonder why did the people not try to optimise ext2 earlier instead of developing ReiserFS. Gruss/Regards, Eduard.
Re: So: reiserfs or ext3 (was Re: ext3 to be in 2.4.15!)
on Sun, Nov 11, 2001 at 11:59:55AM +0100, Eduard Bloch ([EMAIL PROTECTED]) wrote: #include hallo.h Karsten M. Self wrote on Sat Nov 10, 2001 um 02:48:33PM: Now that you said this... I'd like to see how reiserfs and ext3 in writeback mode perform. I'd think ext3 would still be outperformed by reiser for the large dir listings. The issue isn't writing the directories, but searching them. Hash beats linear scan, above the hash overhead threshhold. This crossover's clearly evident for 10,000 directory entries, and is probably significant at substantially fewer. Please first how the thing works before adjucating relying on the traditional prejuducises (like speed: hash tree = list). This ist not allways true if you know how the client program work. In case of ext2, there were allways a problem: when doing file access in a directory (readdir), the ext2 driver did allways start at the top of the list and searched for that entry. Ext3 stores the position the last found entry, so when accessing again, the search begins an this position and you get your entry wery fast, in best case within the first search loop. My understanding is that the difference between ext2 and ext3 is the presence and use of the .journal file. The issue of list access remains, and in particular, random file selects, inserts, and deletes, must operate on the list. I'd be more convinced of your argument if you'd provide a reference to ext3 docs describing the behavior you mention. Peace. -- Karsten M. Self kmself@ix.netcom.com http://kmself.home.netcom.com/ What part of Gestalt don't you understand? Home of the brave http://gestalt-system.sourceforge.net/ Land of the free Free Dmitry! Boycott Adobe! Repeal the DMCA! http://www.freesklyarov.org Geek for Hire http://kmself.home.netcom.com/resume.html pgpcMbBns5ODa.pgp Description: PGP signature
Re: So: reiserfs or ext3 (was Re: ext3 to be in 2.4.15!)
#include hallo.h Karsten M. Self wrote on Sun Nov 11, 2001 um 03:49:33AM: the list and searched for that entry. Ext3 stores the position the last found entry, so when accessing again, the search begins an this position and you get your entry wery fast, in best case within the first search loop. My understanding is that the difference between ext2 and ext3 is the presence and use of the .journal file. The issue of list access Basicaly yes, but they did also tune the driver a bit. remains, and in particular, random file selects, inserts, and deletes, must operate on the list. Of course, the random access does not profit from the trick, but list operations on large directories where ext2 sucked so much. I'd be more convinced of your argument if you'd provide a reference to ext3 docs describing the behavior you mention. Well, it was discussed even on debian-devel. file:///usr/doc/kernel-patch-ext3-2.4/changelog.gz (see 0.9.10) http://lists.debian.org/debian-devel/2001/debian-devel-200108/msg01303.html http://lists.debian.org/debian-devel/2001/debian-devel-200108/msg01238.html Gruss/Regards, Eduard. -- Joey Ok, da steckt auch nicht mehr Arbeit fuer mich hinter? Dann bin ich's doch. -- #debian.de pgpmnpeQhg0Q4.pgp Description: PGP signature
Re: So: reiserfs or ext3 (was Re: ext3 to be in 2.4.15!)
On Saturday, November 10, 2001, at 05:03 PM, Karsten M. Self wrote: on Sat, Nov 10, 2001 at 03:16:54PM -0500, Tom Allison ([EMAIL PROTECTED]) wrote: So, unlike what I read in ReiserFS, I can just hang out on this for a while and then upgrade at my leisure? This is so totally cool Is there an performance difference between this conversion and starting from scratch? I've got both ext3fs and reiserfs on my most recent laptop build. There are advantages to each. I still prefer SGI's XFS. Even the Veritas SE I spoke to a couple of months ago couldn't find anything bad to say about XFS. The only thing he could say was that the Veritas FS is available for Solaris and XFS isn't.
Re: So: reiserfs or ext3 (was Re: ext3 to be in 2.4.15!)
on Sun, Nov 11, 2001 at 04:35:06PM +0100, Eduard Bloch ([EMAIL PROTECTED]) wrote: #include hallo.h Karsten M. Self wrote on Sun Nov 11, 2001 um 03:49:33AM: the list and searched for that entry. Ext3 stores the position the last found entry, so when accessing again, the search begins an this position and you get your entry wery fast, in best case within the first search loop. My understanding is that the difference between ext2 and ext3 is the presence and use of the .journal file. The issue of list access Basicaly yes, but they did also tune the driver a bit. remains, and in particular, random file selects, inserts, and deletes, must operate on the list. Of course, the random access does not profit from the trick, but list operations on large directories where ext2 sucked so much. I'd be more convinced of your argument if you'd provide a reference to ext3 docs describing the behavior you mention. Well, it was discussed even on debian-devel. file:///usr/doc/kernel-patch-ext3-2.4/changelog.gz (see 0.9.10) http://lists.debian.org/debian-devel/2001/debian-devel-200108/msg01303.html http://lists.debian.org/debian-devel/2001/debian-devel-200108/msg01238.html Interesting, thanks, I've read both, will look into it. -- Karsten M. Self kmself@ix.netcom.com http://kmself.home.netcom.com/ What part of Gestalt don't you understand? Home of the brave http://gestalt-system.sourceforge.net/ Land of the free Free Dmitry! Boycott Adobe! Repeal the DMCA! http://www.freesklyarov.org Geek for Hire http://kmself.home.netcom.com/resume.html pgp7P5VlwpPa7.pgp Description: PGP signature
Re: ext3 to be in 2.4.15!
On Sat, Nov 10, 2001 at 11:10:50AM -0600, DvB wrote: ... it's in as of 2.4.15-pre2 http://www.kernel.org/pub/linux/kernel/testing/patch-2.4.15.log Great, they're weren't even done with 2.4.14 yet. Is it just me, or are they churning out releases too damn fast for comfort? I am reassured by Debian's slow release cycle that it will be of high quality. Kernel.org is slowly becoming as bad as any other big company. Mike -- Michael P. Soulier [EMAIL PROTECTED], GnuPG pub key: 5BC8BE08 ...the word HACK is used as a verb to indicate a massive amount of nerd-like effort. -Harley Hahn, A Student's Guide to Unix pgpqOHxsMv1UC.pgp Description: PGP signature
Re: ext3 to be in 2.4.15!
Michael P. Soulier wrote: | Great, they're weren't even done with 2.4.14 yet. Is it just me, | or are they churning out releases too damn fast for comfort? I am | reassured by Debian's slow release cycle that it will be of high | quality. Kernel.org is slowly becoming as bad as any other big | company. It's a conscious choice. I've read that Alan Cox is a big fan of lots of small-change releases rather than a few big ones. As bad as any other big company? How often does Adobe release a new version of Photoshop or Illustrator? -Jason -- Jason [EMAIL PROTECTED] http://wonka.hampshire.edu/~jason It's lucky you're going so slowly, because you're going in the wrong direction. pgp6cv3eqtXH5.pgp Description: PGP signature
Re: ext3 to be in 2.4.15!
* Michael P. Soulier ([EMAIL PROTECTED]) spake thusly: Great, they're weren't even done with 2.4.14 yet. Is it just me, or are they churning out releases too damn fast for comfort? release early, release often. it's easier to locate and correct bugs in small change releases than it is in monster release. -- ) ,_),_) (-(__ |_ _ _ |/ ) | |(_)(_ |\ ( \_, ___ | http://www.exitwound.org : hard to find | ___ | BOFH excuse #419: Repeated reboots of the | | system failed to solve problem| ___ -BEGIN GEEK CODE BLOCK- | Version: 3.1 | | GJ/IT d- s: a C+++$ UL P+++ L+++ E--- W++| | N+@ o K- w O- M- V PS+ PE Y+ PGP++ t+@ 5@ X++ | | R tv+@ b+ DI D+ G++ e h r+++ y+++ | --END GEEK CODE BLOCK--
Re: ext3 to be in 2.4.15!
On Sat, Nov 10, 2001 at 12:53:38PM -0500, Jason Wojciechowski wrote: It's a conscious choice. I've read that Alan Cox is a big fan of lots of small-change releases rather than a few big ones. I guess it makes sense to release that way, I just wish they versioned a little differently, so I could decide what version to upgrade to without incorporating the bug of the week. Mike -- Michael P. Soulier [EMAIL PROTECTED], GnuPG pub key: 5BC8BE08 ...the word HACK is used as a verb to indicate a massive amount of nerd-like effort. -Harley Hahn, A Student's Guide to Unix pgpvvrMzu33Ex.pgp Description: PGP signature
Re: ext3 to be in 2.4.15!
Michael P. Soulier wrote: On Sat, Nov 10, 2001 at 12:53:38PM -0500, Jason Wojciechowski wrote: It's a conscious choice. I've read that Alan Cox is a big fan of lots of small-change releases rather than a few big ones. I guess it makes sense to release that way, I just wish they versioned a little differently, so I could decide what version to upgrade to without incorporating the bug of the week. Mike Well, it's pretty easy. I upgrade when I need to and then I usually either hang back at least one revision or wait ~2 weeks before I do it, to make sure the smoke has a chance to clear. I went from the stable kernel in potatoe to 2.4.9 to 2.4.12. I had to get to 2.4.12 because 2.4.9 had some irda problems. I won't get getting into 2.4.15 for a while yet as I'm curious to see how this ext3 really shakes out and how it's used. After all, wouldn't I have to reinstall everying on a new set up partitions in order to get the support for ext3???
Re: ext3 to be in 2.4.15!
I went from the stable kernel in potatoe to 2.4.9 to 2.4.12. I had to get to 2.4.12 because 2.4.9 had some irda problems. I won't get getting into 2.4.15 for a while yet as I'm curious to see how this ext3 really shakes out and how it's used. 2.4.15-pre2 locks my box hard... After all, wouldn't I have to reinstall everying on a new set up partitions in order to get the support for ext3??? No, you just have to use tune2fs to add a journal file to each partition, and change your fstab. ext3 is forward and backward compatible with ext2. Oh -- you will need a recent version of e2fsprogs J. --
Re: ext3 to be in 2.4.15!
Jeronimo Pellegrini wrote: After all, wouldn't I have to reinstall everying on a new set up partitions in order to get the support for ext3??? No, you just have to use tune2fs to add a journal file to each partition, and change your fstab. ext3 is forward and backward compatible with ext2. Oh -- you will need a recent version of e2fsprogs J. So, unlike what I read in ReiserFS, I can just hang out on this for a while and then upgrade at my leisure? This is so totally cool Is there an performance difference between this conversion and starting from scratch?
Re: ext3 to be in 2.4.15!
On Sat, Nov 10, 2001 at 03:16:54PM -0500, Tom Allison wrote: No, you just have to use tune2fs to add a journal file to each partition, and change your fstab. ext3 is forward and backward compatible with ext2. Oh -- you will need a recent version of e2fsprogs J. So, unlike what I read in ReiserFS, I can just hang out on this for a while and then upgrade at my leisure? This is so totally cool Is there an performance difference between this conversion and starting from scratch? I have converted all my partitions to ext3 quite quickly -- the filesystem format is the same, and tune2fs will jsut add a file to it (it will be called .journal) - then, change fstab, install new kernel with ext3 support, and reboot. See, there's no difference between converting and starting from scratch'. Starting from scratch would mean create a ext2 filesystem and add the journal file to it, etc... (And this is absolutely great!) Read some documentation on ext3: http://people.spoiled.org/jha/ext3-faq.html J. --
Re: ext3 to be in 2.4.15!
Jeronimo Pellegrini wrote: See, there's no difference between converting and starting from scratch'. Starting from scratch would mean create a ext2 filesystem and add the journal file to it, etc... (And this is absolutely great!) Read some documentation on ext3: http://people.spoiled.org/jha/ext3-faq.html J. COOL!!! COOL!!! COOL!!! I have something to PLAY with tomorrow
Re: ext3 to be in 2.4.15!
on Sat, Nov 10, 2001 at 12:37:34PM -0500, Michael P. Soulier ([EMAIL PROTECTED]) wrote: On Sat, Nov 10, 2001 at 11:10:50AM -0600, DvB wrote: ... it's in as of 2.4.15-pre2 http://www.kernel.org/pub/linux/kernel/testing/patch-2.4.15.log Great, they're weren't even done with 2.4.14 yet. Is it just me, or are they churning out releases too damn fast for comfort? I am reassured by Debian's slow release cycle that it will be of high quality. Kernel.org is slowly becoming as bad as any other big company. You clearly don't recall the time of hourly kernel releases ;-) Granted, this was 0.9x days, IIRC. -- Karsten M. Self kmself@ix.netcom.com http://kmself.home.netcom.com/ What part of Gestalt don't you understand? Home of the brave http://gestalt-system.sourceforge.net/ Land of the free Free Dmitry! Boycott Adobe! Repeal the DMCA! http://www.freesklyarov.org Geek for Hire http://kmself.home.netcom.com/resume.html pgp9b7g7yKl8C.pgp Description: PGP signature
So: reiserfs or ext3 (was Re: ext3 to be in 2.4.15!)
on Sat, Nov 10, 2001 at 03:16:54PM -0500, Tom Allison ([EMAIL PROTECTED]) wrote: So, unlike what I read in ReiserFS, I can just hang out on this for a while and then upgrade at my leisure? This is so totally cool Is there an performance difference between this conversion and starting from scratch? I've got both ext3fs and reiserfs on my most recent laptop build. There are advantages to each. Reiserfs has better performance with larger filesystems, particularly for large directory listings. In one case, I've got a directory with 125,000 files in it. Under ext*, directory operations are very slow, on the order of several seconds, due to the need to scan a file list. Reiserfs's use of a hash to store directory entries makes manipulation far faster. OTOH, ext3fs has far less disk overhead for small partitions. Word is that the reiserfs journal requires 32MB, regardless of partition size, while ext3's .journal file is sized proportionately. So, my /, /boot, and /tmp partitions are ext3fs, /var, /usr, and /home are reiserfs. There's room for all. And that's the majick of GNU/Linux: there's choice. It's about choice. Incidentally, those who don't want to wait for the 2.4.15 support can apply the ext3fs patch from Debian. Peace. -- Karsten M. Self kmself@ix.netcom.com http://kmself.home.netcom.com/ What part of Gestalt don't you understand? Home of the brave http://gestalt-system.sourceforge.net/ Land of the free Free Dmitry! Boycott Adobe! Repeal the DMCA! http://www.freesklyarov.org Geek for Hire http://kmself.home.netcom.com/resume.html pgpYpmFL0zOAC.pgp Description: PGP signature
Re: So: reiserfs or ext3 (was Re: ext3 to be in 2.4.15!)
On Sat, Nov 10, 2001 at 02:03:10PM -0800, Karsten M. Self wrote: I've got both ext3fs and reiserfs on my most recent laptop build. There are advantages to each. Reiserfs has better performance with larger filesystems, particularly for large directory listings. In one case, I've got a directory with 125,000 files in it. Under ext*, directory operations are very slow, on the order of several seconds, due to the need to scan a file list. Reiserfs's use of a hash to store directory entries makes manipulation far faster. OTOH, ext3fs has far less disk overhead for small partitions. Word is that the reiserfs journal requires 32MB, regardless of partition size, while ext3's .journal file is sized proportionately. Now that you said this... I'd like to see how reiserfs and ext3 in writeback mode perform. I'd think ext3 would still be outperformed by reiser for the large dir listings. The default for ext3 is ordered data mode, which is safer than writeback. As far as I know, ReiserFS does the equivalent of ext3-writeback only (journalling metadata, but not data). Is this correct? That would bring up another advantage of ext3 (it'd be safer than reiser in ordered data mode) J. --
Re: So: reiserfs or ext3 (was Re: ext3 to be in 2.4.15!)
on Sat, Nov 10, 2001 at 08:12:09PM -0200, Jeronimo Pellegrini ([EMAIL PROTECTED]) wrote: On Sat, Nov 10, 2001 at 02:03:10PM -0800, Karsten M. Self wrote: I've got both ext3fs and reiserfs on my most recent laptop build. There are advantages to each. Reiserfs has better performance with larger filesystems, particularly for large directory listings. In one case, I've got a directory with 125,000 files in it. Under ext*, directory operations are very slow, on the order of several seconds, due to the need to scan a file list. Reiserfs's use of a hash to store directory entries makes manipulation far faster. OTOH, ext3fs has far less disk overhead for small partitions. Word is that the reiserfs journal requires 32MB, regardless of partition size, while ext3's .journal file is sized proportionately. Now that you said this... I'd like to see how reiserfs and ext3 in writeback mode perform. I'd think ext3 would still be outperformed by reiser for the large dir listings. The issue isn't writing the directories, but searching them. Hash beats linear scan, above the hash overhead threshhold. This crossover's clearly evident for 10,000 directory entries, and is probably significant at substantially fewer. -- Karsten M. Self kmself@ix.netcom.com http://kmself.home.netcom.com/ What part of Gestalt don't you understand? Home of the brave http://gestalt-system.sourceforge.net/ Land of the free Free Dmitry! Boycott Adobe! Repeal the DMCA! http://www.freesklyarov.org Geek for Hire http://kmself.home.netcom.com/resume.html pgpMlAKyH62UC.pgp Description: PGP signature
