Cluster Management - Big Data Made Easy: A Working Guide to the Complete Hadoop Toolset - page 232

The cluster that has been installed has a single server rack called “rack 1.” In Figure 8-7 , this is represented by the
five green bars in “Host Disk Space Used %.” Each bar represents a server in the rack. If there were more servers and more
racks, there would be more rows of colored bars. Green represents a good state, while red would represent a problem.
The color-coded key on the left gives the meaning of the color state in the display. In this case, the key presents
disk space usage and it warns when the disk space becomes low on each server.
There is a drop-down menu on the top left of the display so that different Hadoop cluster metrics can be
examined—for instance, memory.
Clicking Services or Hosts in the top bar (next to Dashboard) enables you to see the cluster state from a service or
server point of view. For example, by selecting an individual server, such as the Name Node, in the Hosts interface,
I obtain a detailed view of the state of the Name Node server, as shown in Figure 8-8 .
Figure 8-8. State of Ambari server
Figure 8-8 indicates that the system CPU (blue) on the top left graph is spiking to 100 percent. Also, it shows that
memory usage for processes and disk usage seems to be high. I'm not worried about this, though, as I know that my
cluster is a little under-powered, especially where memory is concerned. However, if you see graphs maxing out like
this on your own cluster, you may need to invest in extra resources or examine the cluster loading.
Clicking Admin in the top menu bar will allow you to access the administration functions within Ambari, thus
enabling you to examine users, high availability, the cluster, security, and access.
Finally, from the Dashboard you can also access user interfaces via links for such components as the Name Node,
Resource Manager, HBase, Storm, Oozie, Ganglia, and Nagios by using the menu on the left of the screen. Many of
these interfaces are familiar because I've covered them in earlier chapters of this topic; it's just that Ambari brings
them all together in one place. For example, Figure 8-9 illustrates how the Resource Manager user interface lists
successful jobs that have completed.