In the example I posted, I'm talking about synthesizing levobupivacaine from scratch. You begin from a chiral starting material (L-lysine) and end up with a chiral product. If you were to start with a racemic mixture of lysine and follow the same procedure, you would end up with a racemic mixture of the two enantiomers of bupivacaine. This approach of planning your drug synthesis so that you obtain only the desired enantiomer at the end is called asymmetric synthesis.
If you have a racemic mixture of bupivacaine and want to isolate only one enantiomer, then you would need some strategy for chiral resolution, the separation of two enantiomers. Chiral resolution is based on the principle that, while enantiomers have identical chemical and physical properties in an achiral environment, their properties differ in chiral environments. In the case of resolving enantiomers using salts, you can take your racemic mixture and react it with the pure enantiomer of a chiral acid or base to form a what is called a diastereomeric salt (basically a salt having at least two chiral centers). Having the chiral acid/base around creates a chiral environment where the two enantiomers of your drug will show different properties (e.g. solubility).
These strategies represent the two main approaches to creating a chiral drug. You can either design an asymmetric synthesis to produce only the desired enantiomer at the end, or you can design a normal synthesis, producing a racemic mixture, and employ a chiral resolution strategy to separate out the two enantiomers at the end.