Sharp on one side but not the other? That's likely a burr.
Burrs are sharp, until they rollover or break off.

Usually you'd get rid of it by alternating passes into the knife with lighter pressure, moving to higher grits.

Although, it might be difficult or awkward to control the direction and pressure with a power tool.
Another thing with some sharpening gizmos is not including the right grits.
Something Iike a 200 grit, then skipping to a 1000 grit can make weird results that kinda feel sharp enough at first, but with a toothier edge where the higher grit cannot really do its job except at the outermost superficial surface and not in the deeper portions left by the low grit.

Honestly? Don't use belt sharpeners, they fuck the metal up at the very edge. I'll explain more if you want, but people turn into assholes about the subject, so I don't go deeper without an invitation to do so.

That being said, what you're dealing with is a burr.

When you sharpen a knife on anything the very edge gets bent over a little. It kinda rolls, and that's what is behind a knife being sharp on one side and not another.

Doesn't matter what the angle is, what the surface is, it's about the physics involved in using abrasives and pressure on steel (or other metals). You can't avoid building a burr, so you don't try to. Instead, you work with it.

Since you're using a grinder, you can't use the usual method of "stropping" where you use the stone itself to gradually thin out and remove the burr (which, at a small enough scale, there's still some burr left).

So you'll have to actually strop, unless you have a stone handy. Any piece of leather will do because you don't need abrasive for this. I say that because they're are traces of silicates in leather, and some hides have more than others, so people tend to think you have țo strop om a leather that's more gritty. All you have to do is alternate running the blade from spine to edge across the leather on each side. This "wiggles" the burr until it snaps off, leaving a more even edge.

Most of the time, people get hung up on the angle being precise and even across the length of the blade and think that's the problem with their results. It very rarely is because the exact angle doesn't matter at all past a fairly obtuse angle (anything where the angle of the edge with both sides taken into account is less than about 45 degrees). It's a little less rare that it's the angle being the same along the edge because as long as the actual edge, the very tippy tip sharp part is even, the rest only matters in terms of how easy it will get through material that's thicker.

Consistent edges help cutting, but they aren't essential to sharpness or keenness.

Sharpness is the degree of angle behind the actual cutting edge, and keenness is how thin that cutting edge is. When you look at edges under a microscope, the difference is more discernable than with the naked eye (and the site knifesteelnerds can show you such images, and links to others). What it amounts to is that it doesn't matter what angle you sharpen at, or what you use to get that angle, what matters is that microscopic edge and how thin it is and how even it is.

So, with a belt grinder, use faster passes across the belt, use mild to soft pressure, and deburr once your angle is set.

You can test for a burr visually, but it's easier to run a fingernail from the spine of the blade down to the edge, on each side. The side with the burr is going to catch the fingernail. But you can't just strop that side. You gotta do both because that burr can switch sides. You actually want it to, because that gets it to come off, leaving the edge as clean as it gets.

Again, don't worry about the edges being even per se. That's only really important in details of use, not in getting a good cutting edge. A good cutting edge is about that microscopic edge being relatively even and as thin as possible.

The blade should always be oriented so the belt is moving in the spine-to-edge direction. Using it in the other direction (into the blade) will grind away the metal faster and less evenly, It sounds like that is what's happening.

It's also worth noting that any kind of inward curve is difficult to sharpen evenly on a belt (or any other wide sharpening surface). You may get better results if you try to work along the edges of the belt instead of the center. For hawkbills and recurves I generally use a narrow sharpener, like an edge of the triangular Spyderco Sharpmaker rods. It takes longer than using a belt, but it produces a very clean bevel and edge.

Best of luck!