On Nov 1, 2007 9:24 PM, Simon Lundell <[EMAIL PROTECTED]> wrote:
> On 10/30/07, Dave Kleikamp <[EMAIL PROTECTED]> wrote:
> > On Tue, 2007-10-30 at 12:35 +0000, Dave Kleikamp wrote:
> > > > That is on a directory that the ordinary 'copy' could not defragment
> > > > more. I had a little trouble aligning buffers and writing files whose
> > > > length was not a multiple of the required alignment.
> > >
> > > I didn't think about those limitations when I suggested it, but that is
> > > something you need to be careful about.  Allocating the buffers with
> > > posix_memalign() should take care of the first concern.
> >
> > As far as the file length, you could pad the file to the next full
> > block, then truncate it to the correct size.
> >
>
> Thanks for the tip. I use a 128Mb statically allocated buffer that is
> reused. The file is written to the next block (512b) boundry and then
> truncated.Seems to work quite well.
>
> The code is available from http://sourceforge.net/projects/freede/. It
> requires boost-1.34. I'll try to figure out how to do a statically linked
> binary.
>
> It works very well for me. I have not lost any files (yet). It checks for
> md5sums before and after the rewrite to se if the file was rewritten OK. Do
> NOT use on directories where files are read from or written to! There is no
> file locking implemented yet.
>
> It currently only tries to minimize the number of fragments per file. It
> does not defragment directories. There was a quite interesting suggestion on
> how to do defrag directories in the thread "jfs tunning", i'll look into
> that next.

Regarding directory fragmentation; is there any way to determine how
fragmented a dir is? Traversal time seems to be pretty useless as the
OS seems to cache directory contents.

//Simon

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
Jfs-discussion mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/jfs-discussion

Reply via email to