Ext3 users - Oct 2001 - Ext3 partition not appearing in df output

Hello,

I converted my / partition to ext3 from ext2 today, and I'm using kernel 
2.4.13+ext3 patch

I used tune2fs -j and set the filesystem to auto in fstab, and compiled ext2 
as well as ext3 in the kernel

Upon rebooting, I tried df -hT to see if it had worked, and to my surprise it 
didn't list my root partition

dmesg say it's fine (and btw everything works fine)

kjournald starting.  Commit interval 5 seconds
EXT3-fs: mounted filesystem with ordered data mode.
VFS: Mounted root (ext3 filesystem) readonly.
EXT3 FS 2.4-0.9.13, 21 Oct 2001 on ide0(3,69), internal journal

I'm using debian unstable
util-linux        2.11l-3           Miscellaneous system utilities.
mount             2.11l-3           Tools for mounting and manipulating 
e2fsprogs         1.25-1            The EXT2 file system utilities and 
libraries.

My partition in on /dev/hdb5

Hello,

I converted my / partition to ext3 from ext2 today, and I'm using kernel
2.4.13+ext3 patch

I used tune2fs -j and set the filesystem to auto in fstab, and compiled ext2
as well as ext3 in the kernel

Upon rebooting, I tried df -hT to see if it had worked, and to my surprise it
didn't list my root partition

dmesg say it's fine (and btw everything works fine)

kjournald starting.  Commit interval 5 seconds
EXT3-fs: mounted filesystem with ordered data mode.
VFS: Mounted root (ext3 filesystem) readonly.
EXT3 FS 2.4-0.9.13, 21 Oct 2001 on ide0(3,69), internal journal

I'm using debian unstable
util-linux        2.11l-3           Miscellaneous system utilities.
mount             2.11l-3           Tools for mounting and manipulating
e2fsprogs         1.25-1            The EXT2 file system utilities and
libraries.

My partition in on /dev/hdb5

----- Original Message -----
From: "dobler" <doblerman@free.fr>
To: <ext3-users@redhat.com>
Sent: Sunday, October 28, 2001 10:38 PM
Subject: Ext3 partition not appearing in df output

> Hello,
>
> I converted my / partition to ext3 from ext2 today, and I'm using
kernel
> 2.4.13+ext3 patch
>
> I used tune2fs -j and set the filesystem to auto in fstab, and compiled
ext2
> as well as ext3 in the kernel
This is propably your problem. When root filesystem is auto, it will also
list this in /etc/mtab, making df think that your root FS is a dummy FS. Try
"df -ha" instead.

This needs to be fixed, as it propably is an Ext3 problem.
>
> Upon rebooting, I tried df -hT to see if it had worked, and to my surprise
it
> didn't list my root partition
>
> dmesg say it's fine (and btw everything works fine)
>
> kjournald starting.  Commit interval 5 seconds
> EXT3-fs: mounted filesystem with ordered data mode.
> VFS: Mounted root (ext3 filesystem) readonly.
^^^^^ this is the problem. The *kernel* mounts the root, and not the
"mount"
program, and /etc/mtab doesn't get updated (correctly).

_____________________________________________________
|  Martin Eriksson <nitrax@giron.wox.org>
|  MSc CSE student, department of Computing Science
|  Umeå University, Sweden

> Im my opinion, mount(8) *should* take care of resolving "auto"
Perhaps, I am not convinced.
Mount has only one real job, that of passing user-specified
parameters to the kernel.

For lazy people some friendliness is built-in, but it is
friendliness that is not guaranteed to work - it is pure
guesswork. Basically an auto mount tries all known filesystem
types in random order until one succeeds. But the call that
succeeds need not at all be "right" - often it is fat or so,
just because the kernel fat filesystem code has so little checking.

So, you cannot maintain that mount "*should*" do any guesswork.
At most that it would be nice. And even there I do not agree.
More and more people do not realize that the guesswork is
guesswork and report a bug if the guess is wrong.
But it is impossible to guess right.
> maintainter of mount?  Is it currently Andreis or not?
Not precisely, but close :-)

Andries

> If you look at it from a big-picture perspective, I suspect
> mount may be a better place to put it than some other alternatives,
> even if this means that we need to make mount "smarter", or
contain
> more filesystem specific gunk.
> (For example, some operating systems
> actually have mount be a dispatch wrapper program like fsck, and all
> of the filesystem-specific options parsing code for NFS, etc., are
> contained in fs-specific programs: mount.nfs, mount.ufs, etc.)
Ted, it seems that you do not understand the problem.
The problem is that if I give you a binary mess, you cannot decide
with certainty what type of filesystem this is supposed to be.
You can only invent some more or less successful heuristics.

If the user refuses to tell mount a type, then mount has to guess.

Andries

> you could just have it dispatch to the right version of /sbin/mount.*
> by looking at the type information in /etc/fstab
Yes, I already thought that you were not understanding the issue.

This discussion started appearing in my mailbox because
someone wrote "auto" in fstab and wanted mount to do something.

Andries

> PS - I added most of the fs types supported by mount
A type that I added a few days ago is sysv
(not yet checked or tested).

#define SYSV_SUPER_MAGIC        0xfd187e20

struct sysv_super_block {
        u_char  s_dummy1[504];
        u_char  s_magic[4];
        u_char  type[4];
};

#define sysvmagic(s)    ((uint) s.s_magic[0] + (((uint) s.s_magic[1]) <<
8) + \
                         (((uint) s.s_magic[2]) << 16) + \
                         (((uint) s.s_magic[3]) << 24))

 * 2001-10-26 Tim Launchbury
 * added sysv magic.

    if (!type) {
            /* sector 1 */
            if (lseek(fd, 512 , SEEK_SET) != 512
                || read(fd, (char *) &svsb, sizeof(svsb)) != sizeof(svsb))
                    goto io_error;
            if (sysvmagic(svsb) == SYSV_SUPER_MAGIC )
                    type = "sysv";
    }

Andries

> Am I right that the kernel _already_ makes use of heuristics
Yes, true. And already it has happened that a filesystem
was destroyed as a consequence.

The correct setup does not involve any guessing.

If you build a large, complicated system, with many layers
of software, and everything is almost always correct, then
the entire system will regularly crash.

So, this guessing of the kernel will go away.
We want to remove guessing, not add more.

Andries

> there is still good reason to want autodetection of udf versus iso9660
that is not at all what the discussion was about

and mount does autodetect a lot of filesystem types