DB2 Subsystem

A DB2 subsystem is one running copy of DB2 on a z/OS system. It is given a short name called the SSID (subsystem identifier), for example DSN, DB2T or DB2P. When someone says "the DB2 subsystem is down", they mean this copy of DB2 is not running.

What makes up a subsystem

• SSID - A 1 to 4 character name that identifies the subsystem on that z/OS image. Programs connect to DB2 using this name.
• Address spaces - The subsystem runs as several z/OS address spaces (DSNMSTR, DSNDBM1, DSNDIST, DSNIRLM) that are started and stopped together.
• Bootstrap dataset (BSDS) - Two small datasets that record which logs are active, where the catalog is, and the subsystem's history.
• Parameter module - Load module DSNZPARM holds the subsystem settings like buffer pool sizes, log sizes and the SSID itself.

Starting and stopping a subsystem

• The subsystem is started and stopped with DB2 commands, usually by the system programmer or DBA from the z/OS console.
• Example:-
  -START DB2 -DISPLAY DB2 -STOP DB2
• -START DB2 brings up all the address spaces. -DISPLAY DB2 confirms the status. -STOP DB2 shuts everything down in the right order.
• A subsystem is normally started once after an IPL and stays up for months. You stop it only for maintenance or a version upgrade.

Data sharing - many subsystems, one data

• In data sharing, several DB2 subsystems (members) on different z/OS images share the same databases. They form a data sharing group with a group name.
• Each member has its own SSID but they all read and write the same tables, with the coupling facility keeping the buffer pools in sync.
• This gives high availability: if one member fails, the others keep serving the same data.