problem with reiser4

problem with reiser4

[fitzboy]

Hello all,
I'm just switching from 2.4/reiser3 to 2.6/reiser4 on the first batch of 
machines. We are running a bunch of apple xRAIDs connected to Dell 
servers running Debian. we have the 2.6.10-mm kernel and we can make the 
reiser4 partitions, but there is a lot of lost space somewhere. If I 
create a reiser3 or xfs partition, I get 1.1T, but with reiser4 I can't 
get above 984GB. That is too much of a jump to be explained by overhead, 
so something is wrong... any thoughts?

Tim

Re: problem with reiser4

[Fred Schaettgen]

On Tuesday 01 March 2005 19:54, fitzboy wrote:
> Hello all,
> I'm just switching from 2.4/reiser3 to 2.6/reiser4 on the first batch of
> machines. We are running a bunch of apple xRAIDs connected to Dell
> servers running Debian. we have the 2.6.10-mm kernel and we can make the
> reiser4 partitions, but there is a lot of lost space somewhere. If I
> create a reiser3 or xfs partition, I get 1.1T, but with reiser4 I can't
> get above 984GB. That is too much of a jump to be explained by overhead,
> so something is wrong... any thoughts?

I'm not an expert, but isn't the free space just a a rough estimation? You 
can't know in advance what the average file size will be etc. Maybe reiser4 
is just more conservative. What matters is how much data is actually stored 
on the disk the moment it runs full.

Fred

-- 
Fred Schaettgen
kde.sch@ttgen.net

Re: problem with reiser4

[Hans Reiser]

Fred Schaettgen wrote:

>On Tuesday 01 March 2005 19:54, fitzboy wrote:
>  
>
>>Hello all,
>>I'm just switching from 2.4/reiser3 to 2.6/reiser4 on the first batch of
>>machines. We are running a bunch of apple xRAIDs connected to Dell
>>servers running Debian. we have the 2.6.10-mm kernel and we can make the
>>reiser4 partitions, but there is a lot of lost space somewhere. If I
>>create a reiser3 or xfs partition, I get 1.1T, but with reiser4 I can't
>>get above 984GB. That is too much of a jump to be explained by overhead,
>>so something is wrong... any thoughts?
>>    
>>
>
>I'm not an expert, but isn't the free space just a a rough estimation? 
>
Uh, no.

>You 
>can't know in advance what the average file size will be etc. Maybe reiser4 
>is just more conservative. What matters is how much data is actually stored 
>on the disk the moment it runs full.
>  
>
It is user error or a bug.  He is saying the reported free space is 
less, and this should not be.  Note that it could be a units of 
measurement issue also....

>Fred
>
>  
>

Re: problem with reiser4

[Łukasz Mierzwa]

Dnia Tue, 01 Mar 2005 10:54:35 -0800, fitzboy <fitzboy@iparadigms.com> napisa³:

> Hello all,
> I'm just switching from 2.4/reiser3 to 2.6/reiser4 on the first batch of
> machines. We are running a bunch of apple xRAIDs connected to Dell
> servers running Debian. we have the 2.6.10-mm kernel and we can make the
> reiser4 partitions, but there is a lot of lost space somewhere. If I
> create a reiser3 or xfs partition, I get 1.1T, but with reiser4 I can't
> get above 984GB. That is too much of a jump to be explained by overhead,
> so something is wrong... any thoughts?

I got the same problem:

gremlin prymitive # mkfs.ext3 /dev/hdd5
mke2fs 1.35 (28-Feb-2004)
Etykieta systemu plików=
Typ OS: Linux
Rozmiar bloku=4096 (log=2)
Rozmiar fragmentu=4096 (log=2)
2562240 i-wêz³ów, 5120710 bloków
256035 bloków (5.00%) zarezerwowanych dla superu¿ytkownika
Pierwszy blok danych=0
157 grup bloków
32768 bloków w grupie, 32768 fragmentów w grupie
16320 i-wêz³ów w grupie
Kopie zapasowe superbloku zapisane w blokach:
         32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632, 2654208,
         4096000

Zapis tablicy i-wêz³ów: zakoñczono
Tworzenie kroniki (8192 bloków): wykonano
Zapis superbloków i podsumowania systemu plików: wykonano

Ten system plików bêdzie automatycznie sprawdzany co ka¿de 24 montowañ
lub co 180 dni, zale¿nie co nast±pi pierwsze. Mo¿na to zmieniæ poprzez
tune2fs -c lub -i.

gremlin mnt # mount -t ext3 /dev/hdd5 /mnt/seagate/
gremlin mnt # df
System plików      bl. 1MB B        u¿yte dostêpne %u¿. zamont. na
/dev/hda1              69790MB   54438MB   15352MB  79% /
none                     262MB       0MB     262MB   0% /dev/shm
/dev/hdd5              20646MB      34MB   19563MB   1% /mnt/seagate
gremlin mnt # umount /mnt/seagate
gremlin mnt # mkfs.reiser4 /dev/hdd5
mkfs.reiser4 1.0.4
Copyright (C) 2001, 2002, 2003, 2004 by Hans Reiser, licensing governed by reiser4progs/COPYING.

Block size 4096 will be used.
Linux 2.6.11-cko2 is detected.
Uuid 74aecbaf-6013-4b9f-9006-691d55f26940 will be used.
Reiser4 is going to be created on /dev/hdd5.
(Yes/No): Yes
Creating reiser4 on /dev/hdd5 ... done
gremlin mnt # mount -t reiser4 /dev/hdd5 /mnt/seagate/
gremlin mnt # df
System plików      bl. 1MB B        u¿yte dostêpne %u¿. zamont. na
/dev/hda1              69790MB   54438MB   15352MB  79% /
none                     262MB       0MB     262MB   0% /dev/shm
/dev/hdd5              19930MB       1MB   19930MB   1% /mnt/seagate

so it's 20646MB on ext3 versus 19930 with reiser4
I used reiser4progs 1.0.4 under 2.6.11-cko2 kernel

£ukasz Mierzwa

Re: problem with reiser4

[Simon Raffeiner]

[-- Attachment #1: Type: text/plain, Size: 3146 bytes --]

One of the 2.6.10-mm kernels killed my Reiser4 Root Partition, yesterday a 
2.6.11-mm one nearly killed a ReiserFS 3.6 partition. Luckily I was able to 
recover it by rebooting with Kanotix and running reiserfsck. Did you try with 
Kernel 2.6.9-mm1, it's the last kernel I am running that runs without 
problems.

Am Sonntag, 13. März 2005 15:46 schrieb Łukasz Mierzwa:
> Dnia Tue, 01 Mar 2005 10:54:35 -0800, fitzboy <fitzboy@iparadigms.com> 
napisał:
> > Hello all,
> > I'm just switching from 2.4/reiser3 to 2.6/reiser4 on the first batch of
> > machines. We are running a bunch of apple xRAIDs connected to Dell
> > servers running Debian. we have the 2.6.10-mm kernel and we can make the
> > reiser4 partitions, but there is a lot of lost space somewhere. If I
> > create a reiser3 or xfs partition, I get 1.1T, but with reiser4 I can't
> > get above 984GB. That is too much of a jump to be explained by overhead,
> > so something is wrong... any thoughts?
>
> I got the same problem:
>
> gremlin prymitive # mkfs.ext3 /dev/hdd5
> mke2fs 1.35 (28-Feb-2004)
> Etykieta systemu plików=
> Typ OS: Linux
> Rozmiar bloku=4096 (log=2)
> Rozmiar fragmentu=4096 (log=2)
> 2562240 i-węzłów, 5120710 bloków
> 256035 bloków (5.00%) zarezerwowanych dla superużytkownika
> Pierwszy blok danych=0
> 157 grup bloków
> 32768 bloków w grupie, 32768 fragmentów w grupie
> 16320 i-węzłów w grupie
> Kopie zapasowe superbloku zapisane w blokach:
>          32768, 98304, 163840, 229376, 294912, 819200, 884736, 1605632,
> 2654208, 4096000
>
> Zapis tablicy i-węzłów: zakończono
> Tworzenie kroniki (8192 bloków): wykonano
> Zapis superbloków i podsumowania systemu plików: wykonano
>
> Ten system plików będzie automatycznie sprawdzany co każde 24 montowań
> lub co 180 dni, zależnie co nastąpi pierwsze. Można to zmienić poprzez
> tune2fs -c lub -i.
>
> gremlin mnt # mount -t ext3 /dev/hdd5 /mnt/seagate/
> gremlin mnt # df
> System plików      bl. 1MB B        użyte dostępne %uż. zamont. na
> /dev/hda1              69790MB   54438MB   15352MB  79% /
> none                     262MB       0MB     262MB   0% /dev/shm
> /dev/hdd5              20646MB      34MB   19563MB   1% /mnt/seagate
> gremlin mnt # umount /mnt/seagate
> gremlin mnt # mkfs.reiser4 /dev/hdd5
> mkfs.reiser4 1.0.4
> Copyright (C) 2001, 2002, 2003, 2004 by Hans Reiser, licensing governed by
> reiser4progs/COPYING.
>
> Block size 4096 will be used.
> Linux 2.6.11-cko2 is detected.
> Uuid 74aecbaf-6013-4b9f-9006-691d55f26940 will be used.
> Reiser4 is going to be created on /dev/hdd5.
> (Yes/No): Yes
> Creating reiser4 on /dev/hdd5 ... done
> gremlin mnt # mount -t reiser4 /dev/hdd5 /mnt/seagate/
> gremlin mnt # df
> System plików      bl. 1MB B        użyte dostępne %uż. zamont. na
> /dev/hda1              69790MB   54438MB   15352MB  79% /
> none                     262MB       0MB     262MB   0% /dev/shm
> /dev/hdd5              19930MB       1MB   19930MB   1% /mnt/seagate
>
> so it's 20646MB on ext3 versus 19930 with reiser4
> I used reiser4progs 1.0.4 under 2.6.11-cko2 kernel
>
> Łukasz Mierzwa

[-- Attachment #2: Type: application/pgp-signature, Size: 189 bytes --]

Re: problem with reiser4

[Łukasz Mierzwa]

Dnia Sun, 13 Mar 2005 15:46:45 +0100, £ukasz Mierzwa <prymitive@o2.pl> napisa³:

> Dnia Tue, 01 Mar 2005 10:54:35 -0800, fitzboy <fitzboy@iparadigms.com> napisa³:
>
>> Hello all,
>> I'm just switching from 2.4/reiser3 to 2.6/reiser4 on the first batch of
>> machines. We are running a bunch of apple xRAIDs connected to Dell
>> servers running Debian. we have the 2.6.10-mm kernel and we can make the
>> reiser4 partitions, but there is a lot of lost space somewhere. If I
>> create a reiser3 or xfs partition, I get 1.1T, but with reiser4 I can't
>> get above 984GB. That is too much of a jump to be explained by overhead,
>> so something is wrong... any thoughts?
>

It's still present in reiser4-for-2.6.11-1 patch:

gremlin prymitive # mkfs.reiserfs /dev/hda6
mkfs.reiserfs 3.6.19 (2003 www.namesys.com)

A pair of credits:
Lycos Europe  (www.lycos-europe.com)  had  a  support  contract  with  us  that
consistently came in just when we would otherwise have missed payroll, and that
they kept doubling every year. Much thanks to them.

Chris Mason wrote the journaling code for V3,  which was enormously more useful
to users than just waiting until  we could create a wandering log filesystem as
Hans would have unwisely done without him.
Jeff Mahoney optimized the bitmap  scanning code for V3,  and performed the big
endian cleanups.


Guessing about desired format.. Kernel 2.6.11-gentoo-r4 is running.
Format 3.6 with standard journal
Count of blocks on the device: 2522192
Number of blocks consumed by mkreiserfs formatting process: 8288
Blocksize: 4096
Hash function used to sort names: "r5"
Journal Size 8193 blocks (first block 18)
Journal Max transaction length 1024
inode generation number: 0
UUID: 5f7b9070-0f62-496c-b288-308663df7689
ATTENTION: YOU SHOULD REBOOT AFTER FDISK!
         ALL DATA WILL BE LOST ON '/dev/hda6'!
Continue (y/n):y
Initializing journal - 0%....20%....40%....60%....80%....100%
Syncing..ok

Tell your friends to use a kernel based on 2.4.18 or later, and especially not a
kernel based on 2.4.9, when you use reiserFS. Have fun.

ReiserFS is successfully created on /dev/hda6.
gremlin prymitive # mount -t reiserfs /dev/hda6 /mnt/cdrw/
df
gremlin prymitive # df
System plików      bl. 1MB B        u¿yte dostêpne %u¿. zamont. na
/dev/hda1               9731MB    5882MB    3849MB  61% /
/dev/hda5              60997MB   32990MB   28008MB  55% /home
none                     262MB       0MB     262MB   0% /dev/shm
/dev/hda6              10331MB      34MB   10297MB   1% /mnt/cdrw
gremlin prymitive # umount /mnt/cdrw
gremlin prymitive # mkfs.reiser4 /dev/hda6
mkfs.reiser4 1.0.4
Copyright (C) 2001, 2002, 2003, 2004 by Hans Reiser, licensing governed by reiser4progs/COPYING.

Block size 4096 will be used.
Linux 2.6.11-gentoo-r4 is detected.
Uuid dfe381a5-0eca-4844-9502-6576da0abc3f will be used.
Reiser4 is going to be created on /dev/hda6.
(Yes/No): Yes
Creating reiser4 on /dev/hda6 ... done
gremlin prymitive # mount -t reiser4 /dev/hda6 /mnt/cdrw/
gremlin prymitive # df
System plików      bl. 1MB B        u¿yte dostêpne %u¿. zamont. na
/dev/hda1               9731MB    5882MB    3849MB  61% /
/dev/hda5              60997MB   32990MB   28008MB  55% /home
none                     262MB       0MB     262MB   0% /dev/shm
/dev/hda6               9817MB       1MB    9816MB   1% /mnt/cdrw

10331MB on reiserfs 3.6 vs 9817MB on reiser4
this was run under amd64 gentoo box
reiser4progs and libaal 1.0.4 were used
kernel was 2.6.11 patched with reiser4-for-2.6.11-1 pulled from namesys server

Can someone help with this bug?

£ukasz Mierzwa

Problem with Reiser4

[Tobi Kunze]

Hi,

i have a strange problem with my system and i think it might be caused by reiser4. If i write a file with vim, even if it is just a word, my harddisk starts to write like mad for a few seconds and vim is not responding. The problem does not exist with nano or emacs. One might now guess it is a vim problem, but it behaves totally normal if i write the file on a fat32 partition on the same disk. Vim just quits and i even cannot hear my harddisk write. Another reason why i think it is a reiser problem is that dpkg got really slow with the last few kernels. It takes very long to extract and install the deb packages.

At the moment i am using Debian and 2.6.14.1 vanilla with reiser4 patch. The problem also exists on 2.6.14-mm2.

With kind regards

Tobias Kunze

Re: Problem with Reiser4

[Artur Makówka]

Tobi Kunze wrote:
> Hi,
> 
> i have a strange problem with my system and i think it might be caused by reiser4. If i write a file with vim, even if it is just a word, my harddisk starts to write like mad for a few seconds and vim is not responding. The problem does not exist with nano or emacs. One might now guess it is a vim problem, but it behaves totally normal if i write the file on a fat32 partition on the same disk. Vim just quits and i even cannot hear my harddisk write. Another reason why i think it is a reiser problem is that dpkg got really slow with the last few kernels. It takes very long to extract and install the deb packages.
> 
> At the moment i am using Debian and 2.6.14.1 vanilla with reiser4 patch. The problem also exists on 2.6.14-mm2.
> 
> With kind regards
> 
> Tobias Kunze
> 
> 

try to remount your reiser4 partition with noatime,nodiratime options, 
it should help.

(it will still be a little slower that what it used to be, but much 
better anyways)