In this paper, we investigate the problem of minimizing average inventory costs subject to a fill rate constraint in a two-stage serial inventory model with a normally distributed demand. Fill rate is the fraction of demand that is satisfied immediately from on-hand inventory. We first find the lower bounds of base-stock levels in each node by using the exact base-stock level that satisfies a fill rate in a single node model proposed by Sobel. And then, we extensively analyze the system and show that the cost function is convex. Using such convexity and some other useful properties, we can easily find optimal base-stock levels numerically.