Database Reference
In-Depth Information
Summary
Hadoop cluster managers will save you time and effort when you want to both install and upgrade your Hadoop
environments. They automatically create the configuration on each cluster server. They also create a set of
configuration files that are more complete and contain fewer errors than the hand-crafted alternative. Also, cluster
managers offer the ability to globally modify your configuration attributes. The savings in time for that feature alone
are huge if you think about manually changing the configuration files on each server in a cluster of even moderate
size. Although I have examined only the Cloudera and Hortonworks cluster managers in this chapter, others are
available from vendors such as MapR.
In short, the monitoring functionality of the Ambari and Cloudera cluster managers is impressive, and having a
single place to check the state of the Hadoop cluster is very useful.
Although it's not a cluster manager, I included Apache Bigtop in this chapter for a number of reasons, but mainly
because it is an attempt to integrate and smoke test a complete Hadoop stack. Cloudera uses Apache Bigtop to create
its CDH stacks, for example. Because the work that the Bigtop project is doing is very worthwhile,I think that Hadoop
users should be aware of it.
Search WWH ::
Custom Search