[clug] Ext2 vs Ext3: weird benchmark result

[Daniel Pittman]

Nathan O'Sullivan <nathan at mammoth.com.au> writes:

>> Also, if it is really mounted as ex3 you should see the kernel ext3 messages
>> in the kernel message log:
>>
>> [71789.271868] kjournald starting.  Commit interval 5 seconds
>> [71789.272296] EXT3 FS on dm-5, internal journal
>> [71789.272310] EXT3-fs: mounted filesystem with ordered data mode.
>
> It doesn't no - that text will appear on a CONFIG_EXT3_FS=y kernel (if
> rootfstype isnt set) as you'd expect.

*nod*

>
> I've done some more testing with a second partition, that confirms what you
> are saying - the root partition is staying as ext2 or ext3 based on whatever
> the kernel starts it as. As you suggest, /proc/mounts shows ext2:
>
> $ cat /proc/mounts | grep ext
> /dev/root / ext2 rw,noatime,nodiratime,errors=remount-ro 0 0

I am honestly pretty shocked that mount will update /etc/mtab to claim ext3,
which sounds very much like a bug to me, but yeah: that matches what I
imagined.

> That was all with Debian 5. Ubuntu 10.04 by contrast seems to report the
> "true" value rather than whats in fstab:

Ah.  They fixed the bug, then.  At least the upstream anticipated my
complaint and resolved it sooner. ;)

[...]

> Thanks for your help, I now feel I understand whats going on.

No problem.  The disparity between /proc/mounts and /etc/mtab (which is where
user-space tracks what is mounted, like fstab tracks what should be mounted)
can be very confusing.

        Daniel

-- 
✣ Daniel Pittman            ✉ daniel at rimspace.net            ☎ +61 401 155 707
               ♽ made with 100 percent post-consumer electrons