O'Reilly logo
  • Murali Parimi thinks this is interesting:

The namenode also knows the datanodes on which all the blocks for a given file are located; however, it does not store block locations persistently, because this information is reconstructed from datanodes when the system starts.

From

Cover of Hadoop: The Definitive Guide, 4th Edition

Note

How to replay the reconstruction of the blocks? what piece of code issues this code issues the reconstruction?