Hello,
http://www.cs.wisc.edu/adsl/Publications/iron-sosp05.pdf
I think I got this link from OLS papers or from fs workshop ... and I really like what it says.
How many of ixt3 features are going to be implemented in ext4?
--
Jure Pečar
http://jure.pecar.org
On Oct 17, 2006 10:36 +0200, Jure Pečar wrote:
> http://www.cs.wisc.edu/adsl/Publications/iron-sosp05.pdf
>
> I think I got this link from OLS papers or from fs workshop ... and I really like what it says.
>
> How many of ixt3 features are going to be implemented in ext4?
None, yet. The journal checksum code has been brought into a basically
"ready-to-go" state, and there is also work to add checksums to the group
descriptors, so there is some chance that this will be added before ext4
is declared "closed".
Nothing done yet to checksum inodes, superblock, file index
or extent blocks. None of the retry code has been looked at.
Cheers, Andreas
--
Andreas Dilger
Principal Software Engineer
Cluster File Systems, Inc.
On Tue, 17 Oct 2006 02:51:30 -0600
Andreas Dilger <[email protected]> wrote:
> None of the retry code has been looked at.
This is what I'm mostly interested in ... predictable, well behaving and tested error handling & recovery. Is ext4 going to be any better here than ext3?
--
Jure Pečar
http://jure.pecar.org
On Oct 17, 2006 11:06 +0200, Jure Pečar wrote:
> On Tue, 17 Oct 2006 02:51:30 -0600
> Andreas Dilger <[email protected]> wrote:
> > None of the retry code has been looked at.
>
> This is what I'm mostly interested in ... predictable, well behaving
> and tested error handling & recovery. Is ext4 going to be any better
> here than ext3?
Not unless someone massages the ixt3 code into a form usable in the
kernel. I'd suggest you contact the authors if you are interested in
doing this work.
Cheers, Andreas
--
Andreas Dilger
Principal Software Engineer
Cluster File Systems, Inc.