Hadoop DFS (cont)
Namenodes ...
- hold file-system meta-data (directory structure, file info)
- e.g. file info: (filename, block#, #replicas, nodes)
- e.g. (/data/a, 1, 2, {1,3}), (/data/a, 2, 2, {4,5}), (/data/a, 3, 2, {3,5})
- provides file open/close/rename operations to clients
- determine replication and mapping of data blocks to DataNodes
- select Datanodes to serve client requests for efficient access
- e.g. node in local rack > node in other rack > remote node
Namenode knows file ok if all relevant Datanodes sent Bockreport
- if not ok, replicate blocks on other Datanodes & update meta-data
|