Master and job nodes explained

A machine running the CLC Science Server software can be considered to be of one of three types:

  1. Single server - a server, to which users submit jobs, and which runs those jobs.
  2. Master node - a machine that accepts jobs from users and then passes them to other systems - either to execution nodes (see below) or a local grid system.
  3. Execution node - here used to describe a machine running the CLC Science Server that accepts jobs directly from a Master node.

Figure 6.2 shows the configuration options for the types of machines running the CLC Science Server.

Image nodetypes
Figure 6.10: The configuration options for the types of machines running the CLC Science Server. The choices of relevance under normal circumstances are SINGLE_SERVER and MASTER_NODE. An administrator will not usually need to manually choose the Execution Node option. This option is there primarily to allow for troubleshooting.