I don't think it solved it/avoided it, so much as throwing more IOPS at the issue means it takes less time.