In FusionInsight HD, when using Streaming's ACK mechanism, which of the following statements are correct? (multiple choice)

Question

Options

• AAfter Acker is enabled, Streaming will identify the Tuple that failed to be sent and automatically
• BAcker will mark the processing timeout or processing failure message as fail.
• CStarting from the spout, the failure of any link in the formed Tuple tree will mark the entire tree as
• DThe application needs to implement message retransmission logic in the fail() interface method of

• A
  29% (6)
• B
  71% (15)

Explanation

Options B, C, and D correctly describe how Storm-based ACK works in FusionInsight HD Streaming. B is correct because the Acker component actively monitors tuple processing and marks any tuple that times out or encounters a processing error as fail. C is correct because the ACK mechanism uses a "tuple tree" model - a spout emits a root tuple that anchors all downstream tuples, so a failure anywhere in that tree propagates the failure status to the entire tree. D is correct because the framework only notifies the spout of failure via the fail() callback; the developer is responsible for writing the actual retransmission logic inside that method.

Option A is wrong because Streaming does not automatically retransmit failed tuples - that is precisely why option D is required. The Acker identifies failures and triggers fail(), but retransmission must be explicitly coded by the application developer.

Memory tip: Think of ACK in three roles - B = Acker is the judge (marks pass/fail), C = one broken link condemns the whole chain (tuple tree), D = the developer must write the rescue plan (fail() logic). If you see "automatically retransmits," that's always a trap - Storm delegates retransmission responsibility to you.

No community discussion yet for this question.