As others have already stated, a lot of the time and effort into Research & Development and modifications on test protocols and exploration on how to fix this issue will be necessary. One thing I would like to bring up is if this type of occurrence was already accounted for, then it would have been considered project slack. Depending on how much slack was considered for this type of failure in this test protocol, then this may not have affected the Gantt Chart all too much. If there was a marginal amount of time planned ahead of time for Research and Development to fix this test failure, then the project Manager could continue with the dependent tasks on the Gantt Chart.

If there was not a lot of slack accounted for then the Project manager will definitely have to reallocate the time into Research & Development and modifying test protocols.