In a sharded replicas set environment with multiple mongos servers, which of the following would decide the mongos failover?

Question

Options

• Amongos
• Bmongo shell
• Cindividual language drivers
• Dmongod

• A
  2% (1)
• B
  11% (5)
• C
  80% (36)
• D
  7% (3)

Explanation

Individual language drivers (C) are responsible for mongos failover because clients connect to mongos instances via a connection string that can list multiple mongos addresses. When one mongos becomes unavailable, the driver detects the connection failure and automatically retries against another mongos in that list - the failover logic lives entirely on the client side.

Why the distractors are wrong:

• A. mongos - mongos is a stateless query router; it has no awareness of or coordination with peer mongos instances, so it cannot orchestrate its own failover.
• B. mongo shell - the shell is just an interactive client tool; while it uses a driver internally, "the shell" is not the mechanism that handles failover in production applications.
• D. mongod - mongod handles replica set elections (failover within a shard), but it has no role in deciding which mongos a client uses next.

Memory tip: Think of it this way - mongos is like a load balancer endpoint; if one goes down, your code (the driver) has to pick another. MongoDB doesn't add a separate "mongos manager" layer, so responsibility falls to the caller - the driver.

No community discussion yet for this question.