Marc Lehmann wrote:
The realtime allocator uses a different binary chop algorithm which while
wasteful, makes it very hard to fragment realtime files. Hmm, buffered works
Hmm, is the realtime code ready for use then? I was under the (likely
wrong) impression that the realtime code is not yet ready.
Let's see if someone answers that, I have no used the buffered over
realtime path, just seen the code.
However, I *guess* I could easily live with, say, 128MB (or even
1GB) realtime extents and live with the internal frgamentation that
will occur: If my interpretation is right (not likely) the realtime
allocator more-or-less treats these realtime extents as the basic unit of
allocation, so internal fragmentation will be extremely high, but external
fragmentation is low?
The realtime extent size is the unit of allocation in the realtime
subvolume,
you still get larger extents than this, they are just always multiples of this.
I don't think you will see fragmentation at all.
Steve
|