We now use the platform is cdh3u4, equivalent to the hadoop 0.2, do not know whether the spark now support? Please give directions,

CodePudding user response:

Compile the source code should be able to, that is, himself, before compiling the source code to modify depend on hadoop package version, to agree with you visit HDFS version, and then compile the package,

CodePudding user response:

Seems not to support it, spark can run on the YARN, can also read the HDFS data, the support is said here that you run on hadoop 0.2