Mail Archives: opendos/1997/04/28/07:29:40
Once upon a time (on 26 Apr 97 at 21:45) Alaric B. Williams said:
> > >You mean /more easily done/.
>
> > Better: You don't have to "lock" files, you don't have race conditions and
> > hundres of other problems... There is far more that can go wrong if your
> > defragging AND allowing write access...
>
> One possible solution is to defrag by rebuilding a copy of the file in a
> linear region somewhere, then switching from the old copy to the new one.
> While a file is being shifted, block read requests come from the old
> version, and block write requests go to the new version, leaving a note to
> the copier not to copy the old version of that block from the original
> contorted file!
>
> The background defrag need not be a perfect defragger. If we can just show
> that it will defragment most files without screwing anything up, then we can
> leave it chuntering away. A full scale bring-the-system-down defrag tool
> will only then be needed if empirical evidence shows that the background one
> isn't good enough!
Hmm... all that seems quite unnecessary. I mean, we can think of (and we have
an example of) a file system with an allocation strategy good enough to make
fragmentation minimal. Having such a system allows us to almost forget about
the defragger - just use it when fragmentation crosses 5% ;-)) (and during
that time we can watch Smurfs or whatever ;-)
==================================================
Stand straight, look me in the eye and say goodbye
Stand straight, we drifted past the point of
reasons why.
Yesterday starts tommorow, tommorow starts today
And the problems seem to be we're picking up the
pieces of a ricochet...
- Raw text -