Database Management Systems

PGA (Program Global Area) contains session-specific memory like variables and arrays when NOT running in MTS (Multi-Threaded Server) mode. In dedicated server mode (non-MTS), each session gets its own PGA. SGA is shared instance memory, SQL Area and Library Cache are components of Shared Pool within SGA - all are shared, not per-session.

Oracle profiles are used to enforce resource limits including idle time. A profile can set IDLE_TIME to limit minutes of inactivity before a session is terminated. Audit tracks actions, quota limits tablespace usage, and 'limit' is not a valid Oracle resource command.

SMON (System Monitor) is responsible for coalescing contiguous free space extents in tablespaces. PMON monitors and cleans failed processes, ARCH archives redo logs, and LGWR writes redo entries. Coalescing improves space utilization by merging adjacent free extents.

DBWR does NOT write to redo log files - that's LGWR's job. DBWR writes dirty buffers from the buffer cache to datafiles. Option A describes correct DBWR behavior (defers writes for batching). Option B describes DBWR's LRU algorithm for cache management. Option D is DBWR's primary function. Writing to redo logs is exclusively LGWR's responsibility.

Option B is false: Oracle guarantees commit confirmation ONLY after the transaction is written to the redo log file. This is fundamental to ACID durability. Options A, C, D are true: one LGWR per instance, long transactions can flush pre-commit (still written), and LGWR is essential for instance survival.

The infodb utility in Ingres is used to check database consistency and integrity. It validates the database structure and reports any inconsistencies or corruption in the database files, making it essential for maintenance and troubleshooting.

The lartool utility in Ingres is used to manually commit transactions. This is particularly useful in recovery scenarios or when automated transaction processing needs to be controlled explicitly by the administrator.

The rollforwarddb command in Ingres is used to restore a database by applying journal files after a restore, bringing it to a consistent state. This command rolls forward the database transactions to make it valid and current. The accessdb utility is for database access management, not restoration.

The accessdb utility in Ingres is used for database user management, including creating users, granting permissions, and managing access controls. It allows administrators to define who can access the database and what operations they can perform. The other options are not standard Ingres utilities for user management.

The unloaddb utility in Ingres is used for data migration by extracting data from database tables into flat files, which can then be loaded into another database. This is the standard utility for data export and migration between Ingres databases. The other options are not valid Ingres migration utilities.

Sybase IQ was the first column-oriented database management system to reach the data warehousing market, launching in the mid-1990s. Its columnar storage architecture was revolutionary for analytical queries and data warehousing workloads. Vertica and ParAccel came later, building on the columnar concept.

Oracle Exadata is the data warehouse database platform developed in cooperation with HP, combining Oracle database software with HP hardware in an engineered system. Exadata uses smart storage scans and columnar compression for high performance. It was introduced as a game-changing data warehouse appliance.

Teradata, Netezza, and DATAllegro are all MPP (Massively Parallel Processing) database systems that distribute data and processing across multiple nodes for scalability. Teradata pioneered MPP architecture, Netezza used it in their appliance, and DATAllegro built an MPP data warehouse platform. Sybase IQ uses a columnar architecture but is not typically classified as MPP.

VOL=SER specifications are required for tape datasets only when they are not cataloged in the system catalog. For cataloged tape datasets, the system can locate the volume through the catalog entry. For uncataloged datasets, you must explicitly specify the volume serial. This is similar to disk datasets - cataloged ones don't need VOL=SER, but uncataloged ones do.

Hierarchical DBMS require navigating through parent-child relationships to access data, while relational DBMS allow non-navigational access using SQL queries that reference data values directly. Both support concurrent access and multiple languages. User access control is not the defining difference.