Availability, Data Sharing, and Operations Management
This domain tests running Db2 12 continuously: data sharing, restart, maintenance, and online operations.
Core rules
- A data sharing group has members sharing a catalog and coupling-facility structures: SCA, lock structure, and group buffer pools. GBP and lock structure sizes must be planned.
- After a member failure, Db2 performs restart, group buffer pool recovery, and retained-lock resolution. Indoubt units of work are resolved by restart or by -RECOVER INDOUBT.
- Online maintenance includes REORG SHRLEVEL CHANGE, online schema changes, and rolling upgrades. Function level activation is group-wide.
- Commands include -START DB2, -STOP DB2 MODE(QUIESCE), -DISPLAY GROUP, -DISPLAY DATABASE, -START DB ACCESS(MAINT), and -DISPLAY THREAD.
- Availability planning includes log placement, BSDS duplexing, CF structure duplexing, and sysplex timers.
Worked example
One member of a three-member group fails. The DBA checks -DISPLAY GROUP, restarts the member, and monitors retained locks. The member performs GBP recovery and resolves indoubt work before accepting new transactions. Reasoning: the group buffer pool may hold changed pages; recovery ensures consistency before the member rejoins.
Traps
- Retained locks can block other members until restart or -RECOVER INDOUBT completes.
- A function level cannot be activated while a member is down at an older level.
- CF structure failure can force a group-wide outage if not duplexed.
- Stopping a member with MODE(FORCE) can leave indoubt units of work.