dfs.block.size has no influence on the performance of TeraGen
I changed the dfs.block.size from 64MB to 1MB in the hdfs-size.xml, and run teragen to explore the influence on job performance. However, the number of map tasks was not changed (I think it should become large, since it is equal with the file size/block size). Also, the time needed to complete the job is almost the same. Is it normal?
TeraGen is for generation of data. So basically this works off no input to generate data for TeraSort.
So basically no. of mappers is not influenced by block size in this case.