This is a plausible theory. But I was seeing very long pauses (> 10 seconds). Is it reasonable that a split takes that long?
I will move to 0.90 today and gather additional time aligned numbers and events to determine if I can see any coincidence. On Wed, Jan 12, 2011 at 8:17 AM, Stack <st...@duboce.net> wrote: > Client loading might be 'random' > spraying over lots of random regions but they all get stuck waiting on > one particular region to come back online. >