1Z0-908 key terms

Shared memory region allocated when an Oracle instance starts, holding the buffer cache, shared pool, redo log buffer, and other structures that every server process accesses concurrently.

Private memory allocated per server process, containing session-specific data such as sort areas, hash join work areas, and cursor state that no other process can read.

Feature where PGA_AGGREGATE_TARGET sets a total memory goal and Oracle dynamically distributes work area memory among sessions, avoiding per-session manual sizing of sort and hash areas.

Capability where SGA_TARGET defines total SGA size and the instance automatically resizes shared pool, buffer cache, and other components based on workload demand, replacing manual per-component tuning.

SGA component that holds copies of data blocks read from datafiles so sessions can read and modify them in memory without physical I/O on every access.

A buffer cache block that has been modified in memory but not yet written to its datafile by the database writer, so its contents differ from what is stored on disk.

Circular SGA area where change vectors describing every modification are staged before the log writer flushes them to the online redo log files.

Set of files recording all changes made to the database, used to replay work during recovery and to support rollback and read consistency, written in cyclic groups.

A copy of a filled online redo log file preserved by the archiver process, forming the basis for media recovery and point-in-time recovery when a datafile is lost.

Database setting in which filled online redo log files are copied to an archive destination before reuse, enabling media recovery and consistent online backups.

Database setting where online redo log files are overwritten without being archived, so recovery is limited to restoring a whole consistent backup and losing changes since then.

Event that advances the checkpoint position, ensuring all modified buffers up to a certain system change number are written to disk and recovery knows where to begin.

Monotonically increasing logical clock Oracle assigns to every change and commit, used to order transactions and to determine exactly which changes recovery must apply.

Background process that writes dirty buffer cache blocks to datafiles, keeping free buffers available and advancing checkpoints without blocking sessions that modify data.

Background process that flushes redo from the log buffer to online redo log files, committing transactions by writing their redo before the commit is acknowledged.

Logical storage container grouping one or more datafiles, used to organize segments and to manage space, quotas, and availability as a single administrative unit.

Operating system file belonging to exactly one tablespace that stores the actual data blocks of the database and can be resized, moved, or taken offline.

Tablespace whose free and used extents are tracked in a bitmap inside its own datafile headers rather than in the data dictionary, reducing contention and recursive space operations.

Contiguous set of data blocks allocated as one unit inside a datafile; a segment grows by acquiring additional extents according to the tablespace's allocation policy.

Database object such as a table, index, or undo segment that consumes storage and consists of one or more extents, possibly spread across multiple datafiles.

Tablespace setting where free space within blocks is tracked by bitmaps, so inserts find room automatically and manual parameters such as PCTFREE tuning become far less critical.

Boundary in a segment marking the highest block ever formatted for use; full table scans read up to this line, so a segment that once grew large stays slow until rebuilt or truncated.

Storage structure holding before-images of changed data, used to roll back uncommitted transactions, provide read consistency, and support flashback queries.

Guarantee that a query sees data as of a single point in time, reconstructing older versions of changed blocks from undo rather than showing uncommitted changes.

Error raised when a long-running query needs an older block image that undo retention has already allowed to be overwritten, typically caused by small undo tablespace or short retention.

Minimum period Oracle keeps committed undo data available, set by UNDO_RETENTION or automatically sized, so long queries and flashback operations can still find needed before-images.

Operation that makes a transaction's changes permanent and visible to other sessions, releasing locks and guaranteeing the redo has been written to the online redo log.

Operation that undoes all changes made by an uncommitted transaction using undo records, returning affected rows and releasing held locks and resources.

Setting that defines how concurrent transactions see each other's changes; Oracle offers read committed and serializable, plus read-only, with readers never blocking writers.

Database design in which one container database hosts many pluggable databases, each with its own schemas and data, sharing one instance and one set of background processes.

The root database in a multitenant configuration that owns the instance, background processes, and shared undo, and contains the root, seed, and all pluggable databases.

Self-contained set of schemas and data that appears as a separate database to applications but shares the instance and resources of its container database.

Account created in the root container whose name and identity are visible in every pluggable database, prefixed with C##, used for administration across containers.

Account that exists only inside a single pluggable database, with privileges scoped to that container and no visibility or login rights from the root.

Right to execute a specific action, granted either as a system privilege such as CREATE TABLE or as an object privilege such as SELECT on a named table.

Named bundle of privileges and other roles that can be granted to users or application code, simplifying administration and enabling dynamic privilege enablement.

Named set of password and resource limits, such as failed login attempts, password lifetime, and idle time, that can be assigned to database users.

Separate process that receives client connection requests on a configured port, then hands each session to a database server process or dispatcher.

Client-side resolution method where a connect identifier is looked up in a tnsnames.ora file that maps the alias to host, port, and service details.

Connection syntax that embeds host, port, and service name directly in the connect string, such as host:1521/orclpdb, requiring no client configuration file.

Logical identifier a client requests when connecting, letting the listener route the session to a specific database or pluggable database regardless of instance name.

Oracle's built-in backup and recovery utility that manages backup sets and image copies, performs incremental and compressed backups, and validates and restores database files.

Separate schema holding metadata about backups, archived logs, and scripts for many databases, enabling longer retention history than the target control file alone.

Small binary file recording database structure, datafile and redo log locations, and the current checkpoint and backup metadata needed to open and recover the database.

Recovery feature that returns the whole database to an earlier point in time by applying flashback logs, without restoring datafiles from backup.

Built-in repository that periodically captures performance statistics and active session samples, producing reports used to diagnose load, waits, and top SQL.

In-memory and persisted sample of sessions currently waiting or running, letting you see which waits dominated a specific time window rather than an average.

Advisory tool that analyzes AWR snapshots, identifies the biggest performance problems, quantifies their impact, and recommends concrete remedies.

Named condition a session is blocked on, such as db file sequential read or enqueue, used to attribute elapsed time to specific bottlenecks during tuning.

Ordered set of operations Oracle chooses to retrieve a result set, showing access paths, join methods, and estimated costs that explain how a SQL statement runs.

Metadata about tables, columns, and indexes such as row counts, distinct values, and histograms that the query optimizer uses to estimate cardinality and choose plans.

Built-in facility for defining jobs, schedules, and programs in the database so recurring work runs at set times, with logging, retries, and resource control.

Server-side export and import utility using expdp and impdp that moves schemas, tables, or whole databases through directory objects, supporting parallelism and remapping.

Utility that loads external flat files into tables using a control file describing the data format, with options for direct-path loading and rejected row handling.

Table whose data lives in an operating system file outside the database; queries read it through the access driver, and it can be loaded with plain SQL.

Graphical and silent tool that installs Oracle software homes, inventories products, and runs prerequisite checks, separate from creating or upgrading a database.

Tool that creates, deletes, and configures databases, including container and pluggable databases, and can generate and run scripts for repeatable deployments.

Command-line utility that applies and rolls back interim patches and patch sets to an Oracle home, checking for conflicts and verifying the inventory.

Directory tree containing installed Oracle binaries and libraries for a specific product version, kept separate from database files and configuration.