Process Architecture

PostgreSQL is client/server RDMS with multi-process architecture and runs on single host. PostgreSQL Server: collection of multiple processes collectively managing one database cluster and it contains following types of processes 1)postgres server process (parent of all processes related to DB cluster management)

• In earlier versions it is called "postmaster"

• It starts up with "pg_ctl" utility with start option

• then it allocates shared memory area in memory

• starts various background processes

• starts replication associated processes and background worker processes if necessary

• waits for connection requests from clients and starts background process whenever connection request is recieved

• Listens to one network port and default is 5432

2)backend processes (handles all queries and statements)

• also called "postgres"

• started by postgres server process

• handles all queries issued by clients

• communicate with client by a single TCP Connection

• terminates when client gets disconnected

• It allow only one database so have to specify database explicitly

• postgreSQL allow multiple clients to connect and "max_connections" controls the maximum number of the clients

• default is 100

3)background process (performs fearture of VACUUM and CHECKPOINT processes) 4)replication associated processes (perform streaming replication) 5)background worker process (perform any processing implemented by users) Memory Architecture: Local Memory Area

• allocated by each backend process

• each area is divided into several sub-areas whos size either fixed or variable

Shared Memory Area

• allocated by postgreSQL server when it starts up

• that area is also divided into several fix sized sub-areas