Database recovery terminology: Is a synchronization point between the database and its transaction log commonly called a “stop point”, or is there a standard term?

Introduction / Context:In database recovery, the system periodically flushes dirty pages and writes log records to establish a known consistent state to speed up crash recovery. This question asks whether such a synchronization point is called a “stop point.”

Given Data / Assumptions:

• We are discussing the standard ARIES-style (or similar) logging and recovery concepts.
• Terminology matters for exam and practical communication.
• The operation establishes a known point in the log from which REDO/UNDO begins.

Concept / Approach:The widely accepted and vendor-agnostic term is checkpoint. A checkpoint records that all transactions up to a certain log sequence number (with some nuances) have their necessary log records on stable storage, and certain pages may be flushed. This reduces recovery time after a crash by limiting how far back the system must scan the log.

Step-by-Step Solution:

Verification / Alternative check:Review documentation of major DBMS (e.g., PostgreSQL, SQL Server, Oracle); all use the term checkpoint for this concept.

Why Other Options Are Wrong:

• “Correct: stop point” uses a nonstandard label and may mislead learners.
• Limiting correctness to specific architectures (distributed or in-memory) does not change the established term.

Common Pitfalls:Confusing checkpoints with savepoints. Savepoints are intra-transaction markers for partial rollbacks, whereas checkpoints concern system-wide recovery state.

Final Answer:Incorrect: the standard term is checkpoint