Modern active distribution networks (ADNs) have become more resilient to service interruptions due to the flexibility offered by integrated distributed energy resources (DERs). Compared to network-owned DERs, directly controlled by the distribution system operator (DSO), a vast majority of DERs controlled by independent virtual power plants (VPPs) are much more difficult to coordinate. The coordination with the VPPs and the management of uncertainty are the two major obstacles in the utilization of these DERs in the ADN restoration (ADNR). This article proposes a hierarchical framework for the sequential ADNR to overcome these two obstacles. The proposed framework adaptively integrates the sequential ADNR planning and the VPP scheduling models based on a comprehensive DSO and VPP coordination scheme to quantify and include the flexibility of VPPs in the ADNR. Furthermore, to address the uncertainty of renewable DERs and loads, a chance-constraint (CC) approach is used in these models. As the probability distributions of the uncertain parameters are not fully known, a distributionally robust reformulation for CCs is proposed that ensures that the solution is robust to any uncertainty distribution defined within a moment-based ambiguity set. The DR reformulation is further linearized to achieve DRCC mixed-integer linear-programming models for ADNR planning and VPP scheduling. Finally, its effectiveness is evaluated on the IEEE 123 node ADN.