No. Most human knowledge is not on the internet. Among the kinds of knowledge not available through the internet are:
- Our myriad individual observations and experiences [example: I know what I ate for breakfast. I know how many pages I read yesterday. I know how many times my houseplant has had white flies.]
- Forgotten knowledge. Scientist David Ehrenfeld has a great essay called "Forgetting," about the significance of knowledge we have forgotten. [his example: Nobody knows how to build certain parts of a battleship turret anymore, he suggests, even though many people knew that only decades ago.]
- Non-propositional knowledge. Many philosophers have argued that we have knowledge when we know how to do things we can't entirely communicate in words [ex: how to play a trumpet well], even if we can record demonstrations of them online or write out instructions about them. Even a full recipe online is typically an incomplete account of how to cook something well.
I have always been struck by Russell Mittermeier's comment that when scientists Ted Parker and Al Gentry were killed in a plane crash, they "carried two-thirds of the unpublished knowledge of Neotropical biodiversity in their minds." I think that immense knowledge is in all three of those categories. So is that I can't remember how to make the toaster at my old house work.
Adding in response to question-edit: I don't think there is a principled way to draw the distinction between knowledge "we learn in school" and things we don't, and I think that is a different distinction than the one between "history and science" and things that aren't history or science. What someone had for breakfast now and then has a place in academic history; what a bird has for breakfast can have an important role in science.
My point is that (1) it's clear that not all scientific knowledge is on the internet, in so far as a lot of scientific knowledge isn't published at all, some has been lost or forgotten, and much remains undigitized or otherwise inaccessible. (2) The very blurry distinction between scientific or historical or otherwise putatively-important knowledge and what doesn't fall in that category means that it's impossible to answer what proportion of it is online, other than "not all."