Choose two. In performance management, which two factors might reduce the ability of an application to scale to a larger number of users?

Question

Options

• Apoorly trained users who do not commit transactions
• Bthe number of tablespaces containing tables updated by a transaction
• Cissuing multiple savepoints during a transaction
• Dthe number of data files containing extents belonging to tables updated by a transaction
• Epoorly written SQL

• A
  79% (38)
• B
  4% (2)
• C
  10% (5)
• E
  6% (3)

Explanation

Poorly trained users who don't commit transactions (A) hold row-level locks and active undo segments for extended periods, creating lock contention that multiplies as concurrent users increase - directly strangling scalability. When a transaction's table extents are spread across a large number of data files (D), each update requires I/O coordination across those files, and under heavy concurrent load this becomes an I/O bottleneck that worsens as users scale. The number of tablespaces (B) is a logical organizational detail that doesn't inherently create contention, and savepoints (C) add negligible overhead since they don't hold additional locks or block other sessions. Poorly written SQL (E) is a common performance concern but is considered a query-level issue rather than a transactional scalability factor in Oracle's performance management framework - the question targets concurrency and I/O infrastructure behavior specifically.

Memory tip: Think "commit or choke" - uncommitted transactions choke locks (A), and too many data files choke I/O (D). Both get worse the more users you add, making them true scalability enemies rather than just local slowdowns.

No community discussion yet for this question.