Tweeting hadoop
Download as PPT, PDF1 like359 views
This document discusses Hadoop and its core components HDFS and MapReduce. It notes that HDFS provides data replication for reliability but failovers are not automatic or physically isolated. It also discusses some of the internal operations of HDFS like create, read, update, delete and configurations related to block size. For MapReduce it lists some key configuration parameters related to task scheduling, maximum tasks and job properties.
Tweeting hadoop
- 1. Tweeting Hadoop Twitter: @zizon
- 2. HDFS Map Reduce etc...
- 3. Disk
- 4. Replication
- 5. But fail overs ... NOT automatically NOT physical isolated
- 6. NFS like It DOSE solve part of it , but not all such as load balance and etc.
- 7. HDFS
- 8. Internals C reate R ead U pdate D elete
- 9. Create
- 10. Create FSNamesystem Inode(IndoeFile,InodeDcitionary, 额头 ) BlockInfo
- 13. Configuration? dfs.block.size io.file.buffer.size dfs.datanode.transferTo.allowed dfs.write.packet.size dfs.datanode.numblocks
- 14. Map Reduce
- 15. Configuration mapred.jobtracker.taskScheduler mapred.tasktracker.map.tasks.maximum mapred.tasktracker.reduce.tasks.maximum mapred.job.split.file mapred.jobtracker.maxtasks.per.job mapred.max.split.size mapred.child.java.opts mapreduce.partitioner.class mapred.reduce.parallel.copies mapred.job.reuse.jvm.num.tasks