Details
Description
In underpowered or loaded environments (like a Docker based virtual cluster), the MR framework may time out a task lagging because chaos has made everything very slow. A simple adjustment to progress reporting can avoid false failures.