No matter if your building a rocket propellant tank, a rocket combustion chamber, or a water bucket, you wind up dealing with the mathematics of tanks. First the concept of volume, for a cylinder; 3.14 * radius squared * length = volume. If you have two hemispherical end caps, the volume of a sphere is 4/3 * 3.14 * radius cubed. Volume times the density of the liquid times the mass of the same volume of water gives you the mass of the liquid that the tank can contain. If the tank is vertical, as most rocket tanks are, or has a tall stand pipe, the pressure (Pounds per Square Inch) of the liquid at the bottom due to gravity will be 0.416 times the height in feet times the density of the liquid. You can also have additional pressure from a pump or trapped gasses. All this pressure must be contained. Fortunately the calculation of hoop stress has been around since the first steam boilers. For PSI, take the diameter in inches, divide by two, because there are two sides, and multiply by the pressure in PSI. Divide this by the thickness of the tank wall, and you know how much stress is in the wall material or PSI * R / T. A tank will also have longitudinal stress. This is the force trying to push the ends off the tank. This stress is PS I * R / 2 * T or about 1/2 of the hoop stress. The hemispheres on the ends will also have stress, the area being 3.14 * radius squared, times the pressure, divided by the circumference or 3.14 * 2 * radius. It turns out that spheres are one half the thickness of cylinders, so spherical tanks are lightest. Unfortunately aerodynamics requires long thin tanks. The end point of all this calculation is that for a given pressure, the mass of the tank material will be proportional to the volume it contains. From a tank's point of view, size makes no difference. Oh, there is one little point about hemispherical Vs ellipsoidal tank ends, but the difference is minimal. FYI the surface area of a cylinder is length * diameter, and a sphere is 4 * 3.14 * radius squared.
Now we get to the important business of hydro testing. A lot of early steam boilers exploded. Then it was found that water is nearly incompressible, and if you get all the air out of a tank and pressurize it with a simple ram pump, you can safely test tanks for their design pressure. I always hydrotest my tanks. Most standard plumbing is designed with a safety factor of 4. Most rockets are designed with a safety factor of 1.5. The hydro testing of rocket propellant tanks is sometimes very expensive, so Arthur Schnitt, yes the Minimum Cost Design guy, came up with the concept of Fracture Resistant Design. Rather than testing every tank, simply trust the design and manufacture of a well tested tank that's a little thicker than needed, and stop testing them. For cheap unmanned rockets and beer cans the system works fine, maybe not for manned vehicles.
And now a little history. The second stage of the Saturn V moon rocket was designed with a very advanced tank system. The contracts were handed out for the first stage and the third stage before the second stage, so all the weight problems of those stages were loaded onto the second stage design. The metal of the tank was sculpted to the precise thickness needed at every point, and no more. When the tank was hydrotested, the Germans in charge of our space program expected the design to have a little extra strength padded into the design. North American Aviation who built the tank told them that there wasn't any extra strength in the design, and so the first tank failed its hydrotest almost exactly at the predicted fail point. The design was accepted and the hydrotest adjusted to a lower level.
To make tankage choices a little easier I offer tankmass.bas which calculates the mass of a steel tank.