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

Re: ext2/ext3 filesystems



On Sun, 28 Dec 2003 17:02:38 -0600, Jeffrey L. Taylor wrote:

> Quoting Jeffrey L. Taylor <jeff@austinblues.dyndns.org>:
>> During the boot, I get the following warning:
>> 
>> Partition check:
>>  /dev/ide/host0/bus0/target0/lun0: [PTBL] [1021/255/63] p1 p2 p3
>> ext3: No journal on filesystem on ide0(3,1)
>> 
>> Is this /dev/hda3 it is talking about?  And why does it think it is an
>> ext3 filesystem?  /etc/fstab and the superblock both think it is
>> ext2.  What thinks it is ext3?
>> 
> 
> AFAICT, this is just an information message.  It is before the partition
> is mounted.  I think because the kernel has ext3 module, it checks
> during the boot process.  I have put ext3 journals on both partitions and
> now I get similar, but opposite message(s) informing me that it found
> a journal.  The partitions are mounted as ext3 filesystems.
> 
> I ran across resources telling/warning that LILO can read ext3
> filesystems, but not all boot loaders can.  For some you have to
> create a tiny ext2 boot partition that the boot loader can read and
> then the other partitions can be anything the kernel, with modules,
> can read (though if the root partition is anything not compiled into
> the kernel, you have to use an initial ramdisk).
> 
> Jeffrey

Jeffrey,

lilo doesn't need to know about filesystems at boot time.  At install time
it creates a map of the files it needs in terms of physical disk
addresses which it uses at boot.  Grub, OTOH, has to grok filesystems at
boot time.  This is why you have to reinstall lilo if you make changes
which invalidate lilo's system map, and also is why lilo is small and
grub is big.

ext3 filesystems can be mounted and read as ext2. The only difference on
the filesystem itself is the presence of the journal. The poster's problem
is the opposite. IMO, He is either trying to mount an ext2 as an ext3, or
the filesystem is ext3 but the journal has been removed.  Or something's
messed with the fs features in the superblock, which seems unlikely to me.

Just an amateur opinion, I'm no filesystem expert.

-- 
....................paul

It's working as coded.




Reply to: