oil paint layers adhesion

This is a complicated question so I will make comments within the body of your query. My responses are in red.
If an oil painting takes years to complete how does an artist get each layer to stick to the next? You hear a lot about fat-over-lean for working in layers but I haven’t heard much about beading up and adhesion problems with working in layers, I have had to work hard to find information.
I have been building up layers of oil paint on canvas and sometimes find that the paint beads up on the previous layer, especially if I make the new layer oily. If I make it less oily and physically brush it around a lot, I can usually get it to finally lay on smoothly, but I am finding some canvases have poor adhesion and after the layer has dried for a few weeks I can easily scrape parts of the new layer of paint off with my fingernail to expose the layer underneath. But not everywhere, just in some areas. This occurs somewhat with a lower surface layer that has dried for one week but more so with a lower surface layer that has dried for a year.
I have generally seen beading when the lower layer is too slick. Simply adding more fatty materials to the next superimposed layer does not solve this and makes the next layer even more slick after it dries. I am guessing that you are adding way too much oil to your paint.
Yes, you can often stop the beading by continuing to brush as the paint sets but in the situation you are describing, you are creating a much less medium rich paint over one that appears to have too much oil media. This is very likely the cause of your poor adhesion. Poor adhesion today can easily lead to delamination tomorrow.
Oil paint that has dried for longer periods of time have gone through much more of their oxidation cycle and are, therefore, less open. It is very common in painting conservation to see later repaints that we can remove from lower layers using a scalpel or even a hard wood pottery tool. The later layers just do not cohere in the same way. Some of this may have too do with absorbed grime or degraded old varnish coatings that remain between the original and later paints, but this does seem to be a recognizable phenomenon.
One of the attractions of adding soft resins like dammar or mastic to paint is that the super sticky and adhesive nature of those resins allowed for the super imposition of numerous layers of fatty media. The problem is that the resins are always going to be very, very sensitive to any solvents brought to the paint, and they become exceeding brittle over time. That is not a major issue if it is in a removable and replaceable surface coating, but it is a huge deal if it is in the paint layers proper.
I haven’t kept track of which oil painting mediums that I used have caused the most problems. I do rinse and dry the surface to remove dust. I don’t want to sand the surface because it seems destructive and would remove part of the painting that I did intend to be there.
Sanding would only be appropriate if it is done on thicker, and more importantly, less precious painted passages. Again, I believe that you are just adding too much medium/oil, etc. to your paint.
So I have quite a few questions about adhesion, if any of you could help me please.
Am I gently scratching too soon, will I find if I wait 6 months that the surface will be harder?
Maybe, but see above.
How durable should a dried oil surface be, should I be able to scrape it with a palette knife and it is fine?
Probably, but see above.
Aren’t I supposed to be using fatter paint for upper layers?
That is probably the cause of problems with many painters. Search for PVC here and forums associated with our moderators. Yes, later layers should not be less flexible than lower layers. The problem is that in practice this is far more difficult to manage. Often it means that painters end up adding increasing larger amounts of fatty media to each successive layer until the whole is so medium rich as to resemble a glassy coating.
What causes the closed off surface? 
Probably the extended oxidative process and the accumulation of grime between layers. I am only writing about oil paint on well made grounds and not the exudation of surfactants and poorly made commercial grounds.

What is the polarity I have heard mentioned?
I am not sure what polarity has to do with this issue. More polar solvents attack oil films more strongly than do lower polarity solvents but even this is a grand overstatement. Toluene is a relatively non-polar solvent that will quickly eat through a youngish oil paint film. There are issues of the difference between the surface activity of a material and that of a coating. That is too complicated for this forum and really has nothing to do with the adhesion issues of common oil paint and mediums. Additionally, I see no positive side of the use of spike lavender oil other than it sort of smells nice…until it is overwhelming.

I understand that there needs to be a mechanical bond and a chemical bond but I don’t quite know how that is failing in this case. If I etch the surface will I get a mechanical bond, but if the surface has oxidized is it too late to get a chemical bond?
It is always best to have both a chemical and mechanical bond. Even that is sort of an oversimplification since how doe you define a chemical bond in this instance?  Going with the common usage, one should suffice but this is less the case with complicated painting techniques involving many layers and the use of too much medium.
I read that you can paint a layer of medium or solvent over the surface to reactivate it, but it hasn’t worked for me. Oil of spike is the strongest solvent I know and it removes well-dried colour but doesn’t seem to ‘etch’ the surface.
Medium (meaning an oil or oil-resin) rubbed over a surface (usually called oiling out) will facilitate the application of an additional layer but it is also fraught with many real problems. This is a huge topic so please read about it in our downloadable pdf entitled VARNISHES in our Resources section.   
https://www.artcons.udel.edu/mitra/resources
One should never etch a painting surface with a solvent, chemical, or vegetable. It will either disrupt the film, weaking it, or leave an extraneous material behind. Onion juice and garlic juice only work to promote whetting out if they remain on the surface.
I have used a method of applying onion juice to lightly etch the surface and I can see that it changes the surface so that water no longer beads up on the surface and it stops the beading of oil paint and the poor adhesion. Is there anything unsound in this technique? I wash the onion off with water.
See above.
I’ve read that using safflower oil makes the poor adhesion of layers worse, is that true?
Safflower oil tends to make a weaker oil paint film than linseed oil. Take that as you will.
Are some pigments worse?
Pigments do affect how strong a paint film is and the speed of drying. One would not want to paint a fast drying, thin lead white containing film over a slow drying red lake. Cracking could ensure. Delamination is of course possible, but I generally think of cracking. Some pigments are effectively anti-oxidants and actively work against the oxidation of the oil paint film. These can also remain far more sensitive to solvents than better drying pigmented films. We can see this discrepancy even after many dozens of years.
Is this surface fragility the same as ‘delamination’?
Not necessarily, but it certainly does not bode well. Think about it like this historical analog. It was very common to paint egg tempera over gold leaf and scratch away the paint to reveal complex designs in 13th-15th c. egg tempera painting. It was called sgraffito. It was a great effect, but the paint had little adhesion until it had dried some time. Even after drying for many, many years, these effects were far more sensitive to mechanical and chemical damage and it is very common to see large sections of sgraffito missing its intended paint even when the rest of the work is in good condition.
Thank you for any help!
Julie