The journaling file system ext3 made it to the test version of 2.4.15 Linux kernel and it seems that it will be as standard of the final version of the kernel. ext3 has pretty much the same design and capabilities of ext2, but it adds journaling. Our Take: Can’t wait for the day that SGI’s next generation, truly advanced file system, XFS, will make it to the main source tree. Update: In a related note, the NTFS driver for Linux was updated today with support for WindowsXP’s NTFS 5 among other new features.
ext3 fs Makes it to the Kernel Source Tree
About The Author
Eugenia Loli
Ex-programmer, ex-editor in chief at OSNews.com, now a visual artist/filmmaker.
Follow me on Twitter @EugeniaLoli
Anyone know if ext3 will make it into BlueOS? And yea, if it does … that probably means the NTFS5 driver will make it too! Very good stuff.
BTW: Anyone know if the updated NTFS driver has experimental support for “writing” to XP partitions? The site doesn’t say.
NTFS5 is transparently tied to the filesystem. It does not care if you have chosen ext2, ext3, JFS, XFS or ReiserFS as your main file system in your Linux partition. It works with all of them.
Personally, I hope and pray that XFS will be the standard filesystem for Linux, with first runner up IBM’s JFS. 😉
AFAIK, the NTFS driver is read-only for security reasons. I think that there was an issue a year ago that the developers should not develop a driver that was able to write to an NTFS partition.
I think what Eugenia is trying to get across is that the Linux NTFS support doesn’t replace support for other filesystems, it just allows you to read your NT partitions. I’m not completely sure because as it’s written it makes no sense at all (I don’t know Greek, but then I don’t run a web site written in Greek, so all’s fair).
The original poster was probably just thinking that if BlueOS play “chase the kernel” as ordinary Linux distributions do, then they will get NTFS5 at the same time as ext3 in their kernel. Frankly it doesn’t much matter, if they branch from the main kernel tree then you’ve got Amithlon again. A neat hack, but ultimately a dead end. If they stick with the mainstream kernel they’re ultimately just another Linux distribution. Damned if you do, damned if you don’t.
I’d be very surprised if Linus announces a new “standard” filesystem for Linux. Some distributions might go the POSIX ACLs route and choose to do that via XFS but I don’t expect the Linus tree will ever express a preference for it. The closest you’re likely to get would be integration of XFS early in 2.5, which still means stable Linus kernels with that feature set are two years away.
If you want XFS for user attributes on Unix, forget it. See all those attributes on a typical live XFS system in the real world? You don’t? That’s because no-one is using them, and it’s quite likely the Unix world will continue to ignore metadata attributes until there’s more shape to the standards for them. Allowing ordinary users to cause anarchy in the namespace is not the Unix way (TM).
>I think what Eugenia is trying to get across is that the Linux NTFS support doesn’t replace support for other filesystems, it just allows you to read your NT partitions.
Exactly. The ext3, is a filesystem used by Linux.
The NTFS for Linux is a filesystem DRIVER for reading Windows NTFS partitions.
>I think what Eugenia is trying to get across is that the Linux NTFS support >doesn’t replace support for other filesystems, it just allows you to read your NT >partitions. I’m not completely sure because as it’s written it makes no sense at >all (I don’t know Greek, but then I don’t run a web site written in Greek, so >all’s fair).
I still don’t know why native english speaking people seems so retarded… the Eugenia posts are fully understandable =P
> Exactly. The ext3, is a filesystem used by Linux.
> The NTFS for Linux is a filesystem DRIVER for reading Windows NTFS partitions.
NTFS is also a filesystem used by linux, just supporst only read-only.
If it can be mounted it’s a filesystem (well it is treated like one, you can mount a mySQL DB too, if you have this virtual filesystem for it
)
kev
“If it can be mounted it’s a filesystem (well it is treated like one, you can mount a mySQL DB too, if you have this virtual filesystem for it
) ”
Now that is interesting – I think ReiserFS’s main hacker said something to the effect that ReiserFS was benchmarked by a web developers’ site as being the world’s fastest web database; and then you have the ancient IBM “filesystems” for their mainframes, which in effect run as databases. And of course, the VM/ESA Shared File System and Byte File System, which are built on top of a relation database and utilize many of its database features.