Hi Steve
On 8 Oct 2002, Steve Lord wrote:
> Performance issues on lookup, or on modifications to the dir? Yes we
> do have hashing in directories, which means lookups do not need to
> walk the whole dir.
>
Good to hear that. I was wondering only on "lookup" case.
> Larger directories are more work to process, and XFS will slowdown
> on creation and removal for large directories, the btree code is
> to blame here.
>
> Steve
I see. Im not a specialist and Ill be very glad if someone would give me a
formula to compute the number of average comparations given the number of
directory entries. "btree" is a binary tree ? whould that mean that
looking up a N entries directory would trigger maximum log2(N)
comparations ?
Also I dont know much about XFS's internal layout, especially transaction
storing on the log. I want to compute the size of a proper log, given the
number of metadata modifications per second (and probably the number of
block writes/reads thad HDD can do in a second). What are the operations
XFS uses journal ? I know of inode modifications which include:
atime/mtime/ctime, size, owner, group, perm.
Thanks
----------------------------
Mihai RUSU
Disclaimer: Any views or opinions presented within this e-mail are solely
those of the author and do not necessarily represent those of any company,
unless otherwise specifically stated.
|