Re: Ahh, that's better...

On Fri, 23 Nov 2007 18:49:13 GMT, Ashikaga <citizenashi@xxxxxxxxxxx>
wrote:

Polychromic of the Cavern #15 howled:
No more TxF meta files taking up space on my drives. I decided to nuke
Vista and use that partition to store my virtual machine harddrives. So
now I have Vista, W95, and XP setup there as Virtual PC machines for when
I need to fiddle with them.

TxF (transactional NTFS journaling) would be nice if it was actually
usable but as it is, it just take us space.

It's a pain removing meta files from the filesystem. Have to use a hex
editor. They really need a tool to do that. Hmmm...I wonder what API
calls exist for manipulating them. Maybe I can write my own tool.

I am learning IT stuff, so I remember journalized file system is to ease IT
auditing. Any of you care to enlighten me on the subject, since my teacher
didn't mention anything specific?

Permissions and disk quotas are for auditing. Journaling is for
preventing data loss or corruption. A meta file called the journal or log
is updated with each impending filesystem operation so that if something
like a power outage interrupts the operation no data is lost.

For example, with a typical non-journaled filesystem like FAT16 when you
move a file the filesystem just writes a change to the file allocation
table, right? If you flipped the power off while it was making those
changes to the FAT, the pointers to the actual file data could be lost.
That's why you'd get those filexxxx.chk files. Those were files where the
pointers had been lost.

With journaling, since the filesystem makes a record ahead of time for
each process, if there is an interruption the filesystem just uses the
last location for the file before the operation. You might have to try
moving the files again after the power comes back on, but the file data
isn't lost.

Transactional NTFS improves the atomic nature of the filesystem operation
supposedly. Have you ever tried copying a folder full of files only to
have Windows balk halfway through, leaving you with no idea which files
got copied? A more atomic journaled operation would mean that if the
operation failed, none of the files would have been copied. It makes the
whole event a pass or fail event so there is no wondering about a
"subevent" where some of the files got copied.

The trouble is that if you're dual booting, Vista converts all the NTFS
drives it sees to use TxF by adding the necessary meta files, which take
up what can be a lot of space and which aren't useable by XP.

Understand?

Vista really ought to ask which drives to convert.
--
The Polychromic Dragon of the -=={UDIC}==-
http://macecil.googlepages.com/index.htm
http://macecil.googlepages.com/safehex.htm
.

Relevant Pages

  • Re: Ahh, thats better...
    ... Vista and use that partition to store my virtual machine harddrives. ... TxF (transactional NTFS journaling) would be nice if it was actually ... It's a pain removing meta files from the filesystem. ... the pointers to the actual file data could be lost. ...
    (rec.games.computer.ultima.dragons)
  • Re: [PATCH 0/4] jbd: possible filesystem corruption fixes
    ... fail to write a metadata buffer to block B in the journal ... reboots and mount the filesystem ... Actually, the journal checksum in ext4/jbd2 detects this kind of error, ... abort journaling ...
    (Linux-Kernel)
  • Re: EXT2 and Windows: WAS Re: FC5 fails to boot after install
    ... support ext3 journaling, but can still read it fine, and can write to it ... Is there a particular issue known when using this driver? ... but I'm more concerned about disk and filesystem ... And one reason to doubt a third-party driver is simply that it isn't ...
    (Fedora)
  • Re: Change Filesystem
    ... The theory (of journaling *filesystem meta-data*) only ... Dances> most of it to an unplanned shutdown. ...
    (comp.os.linux.misc)
  • Re: Simple for Mavericks
    ... I am in no hurry for Vista ... I have heard it is not a part of Vista so no major enhancements to be ... I have never understood why they can not just take Linux filesystem and ... David Liebtag wrote: ...
    (comp.lang.apl)