Grokking the System Design Interview, Volume II
Vote

0% completed

​

Fault Tolerance

Checksums catch corruption in the bytes that survive. Losing a whole DataNode, or losing the NameNode itself, needs a different answer. This lesson looks at how HDFS keeps working through those failures.

How does HDFS handle DataNode failures?

Replication

When a DataNode dies, every block it held becomes unreachable on that node. HDFS protects against this with replication: every block is written to multiple DataNodes, three by default. So when one DataNode goes down, its blocks are still readable from the other replicas.

.....

.....

.....

Like the course? Get enrolled and start learning!

Reading Progress

0%

Vote for new content