AnsweredAssumed Answered

How MapR read file blocks by metadata?

Question asked by li_wei on Oct 12, 2011
Latest reply on Jan 14, 2012 by srivas
In  apache hadoop , namenode records  namespace of all data files(also called metadata), include blocks of files and its location,so  when client read a file from Apache hadoop ,it will firstly query metadata from namenode  and then read block from corresponding data node ,is it right ?

In MapR-cluster, metadatas are distributed into all nodes and metadata is store in containers,when client read a file from MapR-cluster,what process will happen? I guess it will query CLDB at first to get all containers's locations, then scan all containers's metadata to found if this container has this file's data,if yes ,then get data from this container ...

Is my understand right?

Outcomes