The Coding Gopher Courses

1.An interviewer says 'design a photo-sharing app'. What should you do first?

A.Estimate server costsB.Ask clarifying questions and turn the answers into requirements with numbersC.Pick the database technologyD.Draw the load balancer, servers and database

2.A mobile app retries 'Place order' after a dropped response. How do you avoid charging the customer twice?

A.Return 500 so the app stops retryingB.Disable retries in the appC.Use PUT instead of POSTD.Send an idempotency key with each order attempt; the server returns the first result for a repeated key

3.You add a second app server and carts start 'disappearing' at checkout. What's the underlying problem?

A.The second server has a different code versionB.The load balancer is too slowC.The servers keep state (carts in memory); move it to a shared store so any server can handle any requestD.You need a bigger database

4.Two guests book the last hotel room at the same instant and both succeed. Which fix actually prevents it?

A.Add an index on room IDB.Move bookings to a document databaseC.Cache room availabilityD.Lock the room's row on read, or book with a conditional update ('only if still available') and check it changed a row

5.Your cache hit rate goes from 90% to 99%. What happens to database reads?

A.They increase because the cache needs refreshingB.Nothing changes; hit rate only affects latencyC.They drop by a factor of tenD.They drop by about 9%

6.Your database has 10 TB of user data and one machine holds ~2 TB. Will adding read replicas solve it?

A.Yes, if the replicas are synchronousB.No: replication scales reads, not storage or writes. You need sharding (e.g. 5 shards by user ID), each shard replicatedC.No: you need a bigger cacheD.Yes, replicas spread the data across machines

7.What does the CAP theorem actually say?

A.Distributed databases can't be consistentB.Availability always wins in practiceC.During a network partition, a system must choose between consistency and availabilityD.Pick any two of consistency, availability and partition tolerance

8.Your queue uses at-least-once delivery. What must your consumers be?

A.FastB.SynchronousC.Idempotent: processing the same message twice has the same effect as onceD.Ordered

Where are you on the junior → senior path in system design?