Re: [RFC 0/13] extents and 48bit ext3

[Andreas Dilger <adilger@clusterfs.com>]

On Jun 09, 2006  21:43 -0400, Jeff Garzik wrote:
> >???  Can you please be specific in what the performance penalty is, and
> >what specifically is "not sized optimally" after a resize?  How exactly
> >does inode allocation strategy relate to anything at all to online 
> >resizing.
> 
> Inodes per group / inode blocks per group, as I've already stated.

As Stepen and Ted already replied (though I can understand if you missed
it, it seems this is a popular thread :-)- the inode count per group
is a fixed parameter for the whole filesystem that even online resizing
cannot change.

The only things that can change on a per-group basis (with either online or
offline resizing, or with mke2fs -R stride=N, or if there are bad block
on disk) is that the relative offset within the group of the inode and
block bitmaps can change, and the relative location of the inode table
within the group can change.  The size of the inode table per group (and
hence number of inodes per group) is always constant, since it is stored
in the superblock and affects the inode number->group mapping.

Cheers, Andreas
--
Andreas Dilger
Principal Software Engineer
Cluster File Systems, Inc.