MR.Flow is a web-based tool to compose and run workflows of MapReduce modules.

Overview

• easily make a complicated map/reduce flow by drag-and-drop interface.
• use bunch of useful public modules and flows written and published by other users.
• run a flow by just cliking a button and see the running status graphically.

Features

• Composing
  • You can easily write a workflow by connecting MapReduce modules with drag-and-drop user interface.
  • You can register your own MapReduce module and use public modules to compose a workflow.
  • Output of one MapReduce module can be used as input of the next module or the output path can be sent as a JobConf property.
• Execution
  • You can execute a workflow by just clicking a run button without writing a driver code or typing complex commands.
  • MR.Flow automatically generates the driver code and configuration for running and connecting MapReduce modules.
  • MR.Flow executor tab will show you the progress for each module.
  • You can preview the input/output of the modules to check the results are correct.
  • Exceptions thrown by the modules can be seen easily in the module node detail dialog box.
• HDFS browsing
  • Web based HDFS file browser is provided.
  • You can browse files in HDFS and perform basic file system operations like deleting, renaming, moving, uploading and downloading.
• Hadoop clusters
  • MR.Flow provides you a default Hadoop cluster for small jobs.
  • You can register your own Hadoop cluster information and execute your workflows on them.

Screenshot