[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

RE: [OT] Re: Getting ext3 up and running



I referred to the entry

rootfs / rootfs rw 0 0

in the output of "cat /proc/mounts", that is not really very much explicative, due to the setting "auto" in /etc/fstab

and to the fact that file system had has_journal flag and no ext3 module was mounted.
But I was wrong: I think that if you have an ext3 fs and you mount it as ext2, you continue to see the has_journal flag with tune2fs.... I disconnected brain for a while ;-)

I reproduced David situation:

fdisk /dev/hda

created a primary partition big as the whole disk (4Gb)

mkfs -t ext2 /dev/hda1

mount /dev/hda1 /mnt/hda

cp -a /KNOPPIX/* /mnt/hda (it took 11 minutes for 2Gb total)

created /mnt/hda/etc/fstab like

proc       /proc       proc   defaults            0 0
pts        /dev/pts    devpts mode=0622           0 0
/dev/fd0   /mnt/auto/floppy auto   user,noauto,exec,umask=000    0 0
/dev/cdrom /mnt/auto/cdrom  auto   user,noauto,exec,ro 0 0
/dev/hda1 / ext2 defaults 0 0


cp /etc/lilo.conf /mnt/hda/etc

and modified it like this (basic, to improve)

lba32
boot=/dev/hda
prompt
timeout=30
default=Knoppix
image=/vmlinuz
append="lang=us quiet screen=800x600 desktop=xfce BOOT_IMAGE=Knoppix"
root=/dev/hda1
label=Knoppix
read-write

(PS: BOOT_IMAGE flag is parsed by init script in /etc/rcS.d/S00...-autoconfig". if you set it to expert you go into an interactive boot sequence...)

mkdir /mnt/hda/home/knoppix
chown knoppix.knoppix /mnt/hda/home/knoppix


chroot /mnt/hda

lilo

reboot from hd

and fs is ext2 with

mount:
/dev/hda1 on / type ext2 (rw)

then:

reboot from knoppix cd

fsck /dev/hda1  (it was clean)

mount /dev/hda1 /mnt/hda

vi /mnt/hda/etc/fstab

changed fs type for / to auto:

/dev/hda1 / auto defaults 0 0

umount /mnt/hda

reboot from hd

I used tip found at http://mailman.linuxtag.org/pipermail/debian-knoppix/2003-May/002804.html

using the script provided to create an ext2-compressed initrd image instead of cramfs, changing
accordingly /etc/mkinitrd/mkinitrd.conf MKIMAGE variable and running

mkinitrd -o img_file /lib/modules/2.4.22-xfs

I obtain an about 3Mb compressed image file that uncompressed is about 8Mb.
Now you have to strip it to less than 4Mb uncompressed file.

mkdir /tmp/img_big_dir
mkdir /tmp/img_ok_dir
gzip -dc < img_file > /tmp/img_big
mount -o loop /tmp/img_big /tmp/img_big_dir
dd if=/dev/zero of=/tmp/img_ok bs=1024 count=4095
mkfs -t ext2 /tmp/img_ok (and confirm to proceed anyway)
mount -o loop /tmp/img_ok /tmp/img_ok_dir
cd /tmp/img_big_dir 

remove stuff from img_big_dir (tm by Matt ;-)

I removed many fs stuff  and also scsi modules and md modules

when you are under 4Mb

to check dependencies of what you removed

 depmod -a -b /tmp/img_big_dir   

should not give errors or unresolved symbols

then
cd /tmp/img_big_dir 
find . | cpio -pdmauv /tmp/img_ok_dir

depmod -a -b /tmp/img_ok_dir  

umount /tmp/img_ok_dir
gzip -c /tmp/img_ok > /boot/initrd.img
ln -s /boot/initrd.img /initrd.img

add to lilo.conf something like

image=/vmlinuz
initrd=/initrd.img
append="lang=us quiet screen=800x600 desktop=xfce BOOT_IMAGE=Knoppix"
root=/dev/hda1
label=Knoppix-ext3
read-write

run lilo

For me it works, loading journal etc....

with "cat /proc/mounts" I have now

/dev2/root2 / ext3 rw 0 0 


Cheers,
Gianluca










-----Original Message-----
From: ext3-users-admin redhat com [mailto:ext3-users-admin redhat com]On
Behalf Of Juri Haberland
Sent: lunedì 12 gennaio 2004 19.39
Cc: ext3-users redhat com
Subject: [OT] Re: Getting ext3 up and running


[its getting really off-topic now]

Gianluca Cecchi wrote:

> I downloaded knoppix 3.3 and started it inside VMware configured with an ide
> hard disk of 4Gb.

> cat /proc/mounts:
[...]
> /dev/root / ext2 rw 0 0
[...]

> So it seems that ext3 support is compiled into the kernel and all is ok
> with journal.

Where did you get this information from? Just because the fs has the
'has_journal' flag set doesn't mean the kernel is ext3 capable. As you
can see from the /proc/mounts output, the roots is still mounted as ext2.

If you've gone all the way to simulate David's problem, please also be so
kind to try to make a initrd image and boot with it.

Regards,
Juri


_______________________________________________
Ext3-users mailing list
Ext3-users redhat com
https://www.redhat.com/mailman/listinfo/ext3-users




[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]