Server Reintegration in a . . .

A distributed system is a collection of nodes connected by a network, an ideal platform to provide high reliable computing due to the redundancy supplied by a great number of nodes. Node faults and network connection faults can be masked reconfiguring the system. However, sequential faults, that affect multiple nodes can decrease the performance of the system affecting the system reliability and availability. To avoid this, failed nodes should be reintegrated as soon as possible. This paper details the problem of reintegration of failed nodes in a replicated UNIX file system. We built a prototype with the recovery protocols required by the reintegration procedure