On z/OS® systems, the DB2® for z/OS health monitor is started as a task for each DB2 subsystem to be monitored or on a dedicated member of a data sharing group.
The DB2 for z/OS health monitor triggers the evaluation of object maintenance policies at scheduled times and intervals, as defined in the policy. The object maintenance policies are created using the DB2 Control Center's Create Object Maintenance Policy wizard. During each policy evaluation, the criteria for recommending maintenance is checked against the thresholds set in the object maintenance policy to determine the need for object maintenance, that is, whether COPY, REORG, RUNSTATS, STOSPACE, ALTER TABLESPACE, or ALTER INDEX are required, and to identify restricted states, such as CHKP, on table space, index, and storage group objects where applicable. When objects are identified to be in alert state during policy evaluation,the policy health alert contacts are notified at their e-mail addresses or pager numbers. The list of health alert contacts for each DB2 subsystem is defined in and managed from the Control Center.
A snapshot of the evaluation schedule for the policies, which is used by the health monitor to determine when to trigger policy evaluations, is initially taken by the health monitor when it is started. This schedule snapshot is refreshed at the refresh time specified when the health monitor was started, or when the health monitor receives a refresh command. Any change to the evaluation schedule of a policy is picked up by the health monitor when the schedule refresh occurs.
The health monitor is started and stopped from the console, using the MVS™ system START and STOP commands, respectively.
A sample cataloged procedure (DSNHMONP) that starts a DB2 health monitor, and a sample cataloged procedure (DSNHMONA) that starts multiple DB2 health monitors within an MVS system or Parallel Sysplex®, are placed in a procedure library by the installation job DSNTIJHM.
Views, tables, data sets, cataloged procedures, stored procedures, user-defined functions, and the result set table, which are used by the db2 health monitor or the related tasks listed below, are created and installed by the installation jobs DSNTIJCC and DSNTIJHM. DSNTIJCC and DSNTIJHM are shipped with FMIDs JDB771D and JDB881D.
Policy evaluations triggered by the DB2 health monitor are logged in the table DSNACC.HM_EVAL_LOG. An entry is logged when a policy evaluation starts and when a policy evaluation ends. Log entries are kept for 7 days, after which they will be deleted from the table. The DB2 view DSNACC.HM_ALERT_PO_EV, which was created on this table by the DSNTIJCC installation job, can be used to display all policies whose last evaluation iteration was not successful.