Due to the astounding lack of good end-user documentation about Ext4, I've
taken the freedom of writing an article about it for the kernelnewbies
2.6.28 changelog: http://kernelnewbies.org/Ext4
If you find any error or you have some suggestion, just tell me and I'll
fix it.
On Dec 22, 2008 23:58 +0100, Diego Calleja wrote:
> Due to the astounding lack of good end-user documentation about Ext4, I've
> taken the freedom of writing an article about it for the kernelnewbies
> 2.6.28 changelog: http://kernelnewbies.org/Ext4
>
> If you find any error or you have some suggestion, just tell me and I'll
> fix it.
Looks pretty good, thanks for writing this up. Minor notes:
- While > 16TB is supported by ext4 disk formats, the mke2fs/e2fsck code
to support this is not in any released e2fsprogs yet
- I don't think the journal async commit blocks are enabled by default, so
the "20% speedup" is not available yet. There is some work that
needs to be done to make this feature 100% safe in case of corruption.
- I'm not sure if enabling flex_bg on an existing ext3 filesystem is useful
- you don't mention anything about the required e2fsprogs version. I
_think_ (someone should confirm) that e2fsprogs-1.41.3 is the minimum
version that anyone should use.
Cheers, Andreas
--
Andreas Dilger
Sr. Staff Engineer, Lustre Group
Sun Microsystems of Canada, Inc.
Hi, Diego
Thanks for your efforts.
About migrating from ext3 to ext4, you suggested the following:
tune2fs -O extents,uninit_bg,flex_bg,dir_index /dev/yourfilesystem
Tuning on flex_bg isn't supported yet. tune2fs doesn't really do the
magic you want. It just sets the incompatible feature bit in super
block and nothing else. An e2fsck after that won't modify the ondisk
format either. e2fsck doesn't check log_groups_per_flex right
now(might be a bug if superblock corrupts?).
So even if you turn on flex_bg, fs doesn't benefit from flex_bg feature.
On Tue, Dec 23, 2008 at 6:58 AM, Diego Calleja <[email protected]> wrote:
> Due to the astounding lack of good end-user documentation about Ext4, I've
> taken the freedom of writing an article about it for the kernelnewbies
> 2.6.28 changelog: http://kernelnewbies.org/Ext4
>
> If you find any error or you have some suggestion, just tell me and I'll
> fix it.
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
>
--
Cheers,
Bergwolf
................
Margaret Mead - "Always remember that you are absolutely unique. Just
like everyone else."
El Mon, 22 Dec 2008 16:44:56 -0700, Andreas Dilger <[email protected]> escribió:
> Looks pretty good, thanks for writing this up. Minor notes:
> - While > 16TB is supported by ext4 disk formats, the mke2fs/e2fsck code
> to support this is not in any released e2fsprogs yet
> - I don't think the journal async commit blocks are enabled by default, so
> the "20% speedup" is not available yet. There is some work that
> needs to be done to make this feature 100% safe in case of corruption.
> - I'm not sure if enabling flex_bg on an existing ext3 filesystem is useful
> - you don't mention anything about the required e2fsprogs version. I
> _think_ (someone should confirm) that e2fsprogs-1.41.3 is the minimum
> version that anyone should use.
Thanks, I've mentioned all this.