Bill McCormick wrote:
> For dynamic data stored in anywhere in a global having
> globals with compaction around 60% is nice for preventing block splits.
I note that "prevent" is too strong a word.

But is it not correct to say that a KILL compacts a global to some degree?
(Perhaps ISC still discloses the thresholds for this and you can tell us.)

For example, if a KILL leaves two adjacent blocks each 25% full, doesn't 
Cache' combine the data into one block and free the other block?
Or if one block is 10% full and it's neighbor is 40%, are they not usually 
combined?

If so, then would it be correct to say that Cache' is automatically doing 
just about what you suggest?

Reply via email to