What is the purpose of the HBase cluster regularly executing Compaction? (Multiple choice)

Question

Options

• AReduce the number of files under the same ColumnFamily in the same Region
• BReduce the number of files in the same Region
• CImprove data reading performance
• DImprove data writing ability

• A
  72% (36)
• B
  28% (14)

Explanation

HBase Compaction merges multiple HFiles (StoreFiles) within a single Store - and each Store corresponds to exactly one ColumnFamily - so A is correct: the file reduction is scoped to a specific ColumnFamily within a Region, not the Region as a whole. Fewer HFiles per Store means fewer disk seeks during reads, making C correct. D is also correct because when HFile counts grow too large, HBase triggers write-throttling to let compaction catch up; by keeping file counts manageable, Compaction prevents that throttle from engaging and sustains write throughput.

Why B is wrong: It says "files in the same Region" without the ColumnFamily constraint. Compaction operates at the Store level (one ColumnFamily), so it does not reduce files across all ColumnFamilies in a Region simultaneously - the distinction matters for the exam.

Memory tip: Think of Compaction as a janitor for Stores, not Regions - it cleans up one ColumnFamily's file pile at a time (A), making the shelves easier to search (C) and keeping the filing system from getting so cluttered that new deliveries slow down (D). B fails because the janitor doesn't clean the whole building at once.

No community discussion yet for this question.