As we know Default Minimum Container Size is 1024 MBs Consider Hadoop Cluster <1-Master & 3-Datanodes> Namenode : master (3GB RAM, 2 Cores) Datanodes : data1 (512MB RAM, 1 Core), data2(512MB RAM, 1 Core), data3(1GB RAM, 1 Core). Input File Size : "File"=500MB with Replication factor 3 (All the nodes having all the blocks) Default Block Size = 128 MB therefor "File" will be split into 4-Blocks Due to Repli.Factor = 3, all the nodes will be having all the 4-Blocks i.e 4-Containers. Default Minimum Container Size is 1024 MBs Question1 : On which datanode AM will allocate Container? Question2 : Data1 & Data2 has 512MB of RAM, Does that mean AM will never allocate container on these datanodes??? ... View more

As we know Default Minimum Container Size is 1024MBs Consider this Scenario : I have setup a Hadoop Cluster <1-Master & 3-Datanodes> in VIRTUAL BOX. Namenode : master (3GB RAM) Datanodes : data1 (512MB RAM), data2(512MB RAM), data3(1GB RAM). Input File Size : 500MB with Replication factor 3 (All the nodes having all the blocks) QUESTION : 1. Now lets say i want to run a Mapreduce Program, Will containers get allocated on data1, data2 ???? According to my results Job created containers on all three datanodes. 2. Is there any way we can specify a Job to run on specific node??????? ... View more