My guess is the cadmium bromide is causing the film to lift. The albumen subbing should take care of it.
Buy a whetstone or obtain some coarse sandpaper to roughen the side of the glass plate. Just run the whetstone (or sandpaper) along the edge of the plate at about a 45-degree angle about three or four times per edge. You are sanding down the sharp edge of the glass (i.e., the corners running width- and lengthwise between the plate's horizontal surface and the thin vertical side). Make sure to get the corners and do all eight edges. You should wear a respirator mask when doing this because fine pieces of glass will be chipping off the plate and you don't want to breathe those in. (Silicosis is a nasty disease caused by prolonged exposure to silica dust.) Goggles/glasses are probably a good idea too.
Once the edge has been prepared, take the white of an egg and put it in a small bowl. Have a bit of distilled water in another bowl (or bottle cap). Wet a cotton swab ("Q-tip") with the water then dip it into the albumen. Take the swab and hold it against the edge of the glass plate. Run it down the edge so the cotton comes in about 1/8" onto the surface of the plate. You will essentially be painting the edge of the glass with egg white. Go around the perimeter of the plate using additional albumen as needed. Let the albumen dry. I usually prepare a large batch of plates ahead of time, first cleaning and then subbing them, and then I store them for later use. I also typically mark the backside of the plate with a china marker because the albumen is very difficult to see once dried.
Another approach is to sub the entire surface of the plate with albumen (or some chemical made for this purpose like a silane compound- Bostick and Sullivan sell one, I think). Take the white of an egg and beat it thoroughly into a half-liter of distilled water. Filter the that solution to remove any froth and store it in a jar. Pour the solution on the glass plate as if you were pouring collodion and then drain it back to the jar. Dry the plate on a paper towel while propping the plate subbed side down and slightly off vertical against a wall or box. This will help prevent dust from settling into the albumen.
Good luck with the latter approach. The edge-subbing approach is quick-and-easy, much less messy, and dust is not a problem. I much prefer it to subbing the entire plate surface.