The Hadoop Name Node Layer

The tests mapped to this layer:

• Capture corrupt and missing blocks;
• Measure how quickly the Secondary NameNode downloads EditLogs from the NameNode, and how quickly the NameNode uploads and downloads FsImages from the Secondary NameNode;
• Warn administrators of slowness in Journal transactions that may potentially impact cluster availability;
• Alert you to the poor usage of the retry cache;
• Identify the exact RPC interface that is overloaded with connections, and proactively detect RPC processing bottlenecks on the NameNode;
• Quickly detect unscheduled reboots of the NameNode;
• Track administrator- and user-initiated snapshot operations on the HDFS;
• Monitor how storage space is used across a cluster, and be proactively alerted if the cluster is running out of space