Can anyone tell me if read ahead in Lustre includes "early return" features. I mean that if I read 4K and readahead decides to fetch 1M will my request get serviced when the first 4K arrives? Is this important? - Peter -