We study the problem of scheduling jobs on parallel machines minimizing the total completion time, with each job using exactly one resource. First, we derive fundamental properties of the problem and show that the problem is polynomially solvable if $p_j = 1$. Then we look at a variant of the shortest processing time rule as an approximation algorithm for the problem and show that it gives at least a $(2-\frac{1}{m})$-approximation. Subsequently, we show that, although the complexity of the problem remains open, three related problems are $\mathcal{NP}$-hard. In the first problem, every resource also has a subset of machines on which it can be used. In the second problem, once a resource has been used on a machine it cannot be used on any other machine, hence all jobs using the same resource need to be scheduled on the same machine. In the third problem, every job needs exactly two resources instead of just one.