Regarding orbits, your condition is possible. For this, you want your moon to revolve around the planet with a period of exactly half a year. A depiction follows:
You want your planet have a solar eclipse in a certain time of year. Let's position your solar system so that your planet is on the left of its sun, so the planet, its moon and the sun would be positioned like this:
Then, you want the moon to phase in/out for the entire year. Let's check where the moon should be to change phase to 180 degrees in half a year. It will be a full moon, so it should be about behind the planet when viewed from the sun, and the planet itself would complete half a round of its own orbit, now it will be on the right from the sun. Like this:
Notice, the moon's position relative to the planet has not changed. Should we use a quarter period (sorry no hand drawings here), the configuration from the same view would be either like this:
^ (to the sun)
Or like this:
^ (to the sun)
The first variant means that the moon either does not revolve around the planet, or does 4*X revolutions around it over a year. The non-revolving moon is not physically possible, since it will plain fall to the planet making a big BOOM, the 4*X variant does not satisfy the requirement of changing phases over a year, so that variant is out. The second variant means that the moon did 1/2+X (X is integer) rounds around the planet during the planet's 1/4 years. Solving with both conditions to be satisfies yields X=0, thus the moon's orbital period is (1/4)/(1/2) years, or half a year.
The thing that would be impossible is a solar eclipse, or at least a full solar eclipse, if your moon is just a moon. For example, here on Earth we have a full solar eclipse because our Moon is pretty close to the planet, having its angular diameter about the same as the Sun when viewed from the planet's surface. Should we position the Moon far enough so that its orbital period would be 1/2 of a year (183 days instead of 27.32), it will have its orbital radius of
384.4*(183/27.32)^(2/3) = 1365.9 thousand km, and its angular radius when viewed from Earth would decrease
(183/27.32)^(2/3) = 3.553 times, making the Moon's disc cover only 1/(3.553^2) = 0.079 or 7.9% of the Sun's disc. Still plausible to be detected from the planet, though. And for our Earth, this distance is still within its Hill sphere radius which is about 10% larger, so the moon at that distance will still be Earth's satellite, but its orbit would be noticeably affected by the sun, thus a minor correction downwards could be made to the estimated radius in order to retain the moon's period of exactly half a year.
If you would want a moon to be of enough size to cover the entire sun's disc, it will have its radius to be 3.553 times larger, and its mass about
3.553^3 = 44.87 times bigger, this is no longer a moon but another planet, a tad larger than half of Earth. So your planet is no longer a single entity but two planets revolving around their common barycenter at a period of half a year, eclipsing their common sun from each other once a year. Probably that other planet also has an atmosphere, some life on its surface, and there's a possibility for interplanetary travel for either side... Ohhh the possibilities!
So to summarize. Yes it's possible, for that you want to have your planet to be a double, revolving at period of half a year around common barycenter. The masses of counterparts should likely be divided as half-and-half, or a rough estimate of that. There would be a likely problem of tidally locking the planets to each other, or at least some serious tides because the "moon" would be pretty heavy, yet these can be estimated in another question.