Greenplum Unified Analytics Platform (UAP) - Getting Started with Greenplum for Big Data Analytics - page 45

Master host core functions are listed as follows:
Consumption of the parse tree and production of the query plan (query plan
contains how the query is executed, for example, hash join versus merge
join)
Communication of the query plan to segments
Allocation of cluster resources required to perform the job and accumulating/
presenting final results
Note
A query executor (worker process) is responsible for completing its por-
tion of work and communicating its intermediate results to the other work-
er processes. For each slice of the query plan, there is at least one work-
er process assigned. A worker process works on its assigned portion of
the query plan independently. During query execution, each segment will
have a number of worker processes working on the query in parallel.
Relatedworkerprocessesacrosssegmentsthatareworkingonthesame
portion of the query plan are referred to as gangs .
• Standby master : The standby master is a warm standby server that is activ-
ated when the master host is unavailable.
• Node or segment host : Node is the server that has the actual installation of
Greenplum Database (primary and mirror segments instances).
• Each segment holds a portion of data for each distributed table and
index.
• Every segment server can hold multiple segment instances residing
on it. By default, there would be one primary segment and if mirroring
is enabled, it would have one or more mirrors. By definition, number
of primary and mirror segments equal to the number of physical
cores.
• Segment servers are not directly accessed by users; all communica-
tions with the segments are through the master. Another way to con-
nect is in utility mode (a rare case).