Re: EXT2 and Windows: WAS Re: FC5 fails to boot after install

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

 



Dan wrote:
> You can also install the ext2 driver at http://www.fs-driver.org . It doesn't 
> support ext3 journaling, but can still read it fine, and can write to it 
> without the journaling. If you don't have any specific 
> reason to write to your linux partitions, I'd just leave the write support 
> off, since Windows can muck up the filesystems, especially if it 

Mike McCarty wrote:
> Erm? Every OS can muck up a filesystem, especially when it crashes.
> Is there a particular issue known when using this driver? Or what?
> What are you warning against?

Driver bugs. Lack of support for ext3 journalling. And which OS is most
likely to crash, anyway? [1]

I don't know about you, but I'm more concerned about disk and filesystem
integrity than nearly any other part of the OS. An OS that crashes
*without* mucking up the filesystem is much less of a problem than one
which loses data.

And one reason to doubt a third-party driver is simply that it isn't
that widely used, and you might not want to be the person that finds the
next bug. Which is a lot more likely if you're one in (say) ten thousand
users than if you're one in (say) ten million. (Figures plucked from
thin air).

Next point: when Linux closes down cleanly, ext3 filesystems are in a
consistent state and can reasonably be read as ext2. If Linux crashes,
or the power fails, you need data in the journal to get the filesystem
into a consistent shape.

A driver that doesn't write to the filesystem may not get a 100%
consistent filesystem, but on next boot into Linux, fsck will get it
into shape. But if an ext2-only driver boots an inconsistent filesystem
in write mode, it will throw away the data in the journal which could
have been used to recover filesystem data that was being written when
Linux stopped abnormally.

James.

[1] In my experience, that question is closely related to "which OS has
more kernel-mode driver code written by third parties to which the OS
vendor doesn't have source".

Both Linux and Microsoft techs, as far as I can tell, indicate that core
OS code is of better quality than drivers from the same tree (on
average), which are better -- sometimes *much* better -- than
third-party drivers (which might not support such exotic features as
dual-core processors).

This is one reason why Microsoft is trying to get all drivers checked
(by them) and signed.

-- 
E-mail address: james | "Right lads, we've got 45 minutes to score 37 goals.
@westexe.demon.co.uk  | No problem with that -- the other team just did."


[Index of Archives]     [Current Fedora Users]     [Fedora Desktop]     [Fedora SELinux]     [Yosemite News]     [Yosemite Photos]     [KDE Users]     [Fedora Tools]     [Fedora Docs]

  Powered by Linux