Provided are methods, apparatus and computer programs for recovery from failures affecting a server in a data processing environment in which a set of servers controls a client's access to a set of resource instances. Independent of any server failure, the client or a gateway is provided with an identification of both a primary server for accessing the resource and at least one secondary server for use as a backup server for accessing the same resource instance (for example, the same physical storage disk). The client or gateway connects to the primary server to perform resource access operations. Following a failure that affects availability of the primary server, the client or gateway connects to the previously identified secondary server to access the same resource instance. Provision of the identification of at least one backup secondary server (without requiring the ‘trigger’ of a failure) avoids the need to discover a new server as part of the recovery operation following a failure. Release of existing reservations using a reset operation, and re-reservation by the original initiator via a backup server, deals with any dangling reservations.