Your question cannot be answered without some specification of what knowledge is--what counts as knowledge. This topic is extremely controversial among epistemologists. But I think one aspect of your question allows at least a part of an answer to it.
Epistemologists may not agree on the entire analysis of knowledge, but most agree that whatever is known must be true, and most agree that in order to know something you at least have to believe it. The real controversies tend to begin when epistemologists debate what is now often called the "warrant" condition, which is the purposely vague expression used to denote whatever else is needed for knowledge, other than true belief--or to put it slightly differently, whatever it is that distinguished knowledge from other species of true belief.
Now think a little bit about the (relatively uncontroversial) belief condition. What does it mean to believe something? One thing belief is often supposed to include is a dispositional component. Part of what it means for me to believe there is a truck coming down the road towards me is that I am disposed to step off the road surface to get out of its way. If I am not so disposed (assuming I am not seeking to be killed or injured by the truck), then we might wonder if I really believe the truck is coming at me.
Your case tempts me to respond that even though the person has received the information that his father is dead, he does not yet believe it, since at least some of the dispositions in accordance with which we would expect him to act in certain ways appear not (yet) to be present. It is one thing to have (access to) certain information, and another thing actually to believe that information. Being in a state of denial seems to me to be an example of at least impaired cognitive function at the level of belief. One can't know something without believing it, at least in the dispositional sense. So in your case, it looks like it can't count as knowledge until, as you put it, it "hits him."
Whether or not this is really the correct diagnosis of the case, however, will depend upon just how much we build into our account of belief in terms of relevant dispositions. That seems to me a matter of likely controversy, and so it could be that another place to attack this case would be in terms of the warrant condition. As I said, there are lots and lots of different accounts of what warrant consists in (for just a few examples: being completely justified, having one's belief generated or sustained by reliable true-belief-forming processses, or having the belief produced by reliable cognitive processes that are functioning properly within an environment to which they are well suited). One might say that the person's justification cannot be complete unless and until the person recognizes all of the evidence as evidence, and perhaps the initial under-reaction shows that he has not yet achieved that level of justification. Or perhaps the person's cognitive functions are not fully adequate (do not count as functioning properly) until their representation of the fact to him are sufficient to qualify as "hitting him."
Just to muddy the waters a little further, it also seems quite possible to me that someone can know something at a different time than he or she manages to respond to what he or she knows at an emotional level. Just because I am stunned into a kind of non-response to something at first does not necessarily indicate a lack of knowledge; it might instead indicate a lack of readiness to respond or react to what I know at the emotional level. So I think this is also another way to see your case.
I guess if you want the gist of my response in a nutshell, it seems to me that the case is somewhat underdetermined, as presented, which is why it seems to me there are several reasonable reasonses that can be made to it.