yoklainterview sim
ProgressHistory

Database Mid Interview Questions

2066 verified Database Mid interview questions — solve with answers, learn from explanations, test yourself in a real simulation.

Try the real simulation →

Sample questions

Backup RecoveryDifficulty 2
A team runs a primary database with a synchronous replica and argues: 'We don't need backups — the replica is our backup.' Why is this reasoning dangerous?
  • aA replica only serves reads, so you can never copy the lost data back out of it
  • bAn accidental DELETE or corruption is replicated to the standby too, destroying both copies✓
  • cReplication lags so far behind the primary that the standby never holds usable recent data
  • dA replica keeps its data in a compressed on-disk format that cannot be turned back into a database
Explanation:Replication copies changes faithfully — including destructive ones. A bad DELETE, an UPDATE with the wrong WHERE, or corruption propagates to the replica in seconds, so both copies are lost together. A backup is a point-in-time copy you can go back to; a replica is not. Reads (a), lag (c), and storage format (d) are not the core reason.
Backup RecoveryDifficulty 2
How do incremental and differential backups differ?
  • aThey are simply two different names for one and the same backup technique
  • bAn incremental copies everything since the last full backup; a differential copies only the very newest changes
  • cAn incremental may only follow a full backup, while a differential may only follow another differential
  • dAn incremental copies changes since the last backup of any kind; a differential, since the last full backup✓
Explanation:An incremental backup captures only what changed since the previous backup (full or incremental), so a restore must replay the full plus every incremental in order. A differential captures everything changed since the last full backup, so a restore needs only the full plus the latest differential. Option (b) swaps the two definitions.
Backup RecoveryDifficulty 2
Compared with taking a fresh full backup every night, a 'weekly full + nightly incremental' strategy mainly trades...
  • aFaster, smaller nightly backups for a slower restore that replays the full plus each incremental✓
  • bSlower nightly backups for an absolute guarantee that no committed data can ever be lost
  • cExtra storage consumption for the ability to recover the database to any individual second
  • dA simpler setup for the database engine automatically copying each backup to an off-site location
Explanation:Incrementals write far less each night (faster, smaller), but recovery must start from the last full and replay every incremental in sequence — more steps, more time, and more chance that one link is missing. It does not guarantee zero loss (b) or second-level recovery (c); those are separate concerns.
Backup RecoveryDifficulty 2
pg_dump mydb > mydb.sql

This command produces a logical backup. How does a logical backup differ from a physical backup?
  • aA logical backup copies the raw on-disk files and blocks; a physical backup exports SQL statements
  • bA logical backup exports data as SQL or a portable format; a physical backup copies the raw files✓
  • cA logical backup can only capture the schema definition and never the actual rows of table data
  • dA physical backup can run only while the database currently has zero open client connections
Explanation:A logical backup (like pg_dump / mysqldump) reconstructs data as portable statements, so it can restore into a different version or platform. A physical backup copies the raw data files and blocks — faster for huge databases but tied to the same engine layout. Option (a) reverses the two.
Backup RecoveryDifficulty 3
You must move a database to a server running a different major version on a different CPU architecture.
# option A (physical): copy the raw data files
# option B (logical): pg_dump olddb | psql -h newhost newdb

Which is generally the safer choice, and why?
  • aThe physical file copy, since raw data files are always compatible across versions and CPU architectures
  • bNeither works — data moves between servers only by configuring streaming replication first
  • cEither is equivalent here, because the backup format has no bearing on cross-platform portability
  • dThe logical SQL dump, since it is portable and can be reloaded into a different version or platform✓
Explanation:Physical backups copy engine-internal file formats that can differ between major versions and architectures, so restoring them elsewhere is fragile. A logical dump is engine-readable SQL/data, which is exactly why it is the standard tool for cross-version and cross-platform moves. Replication (b) keeps copies in sync but is not a prerequisite for migration.
Backup RecoveryDifficulty 2
The business says: 'In the worst case we can tolerate losing at most 5 minutes of data.' Which requirement does this statement set?
  • aEvery scheduled backup job must finish running within a 5-minute execution window
  • bAfter a failure the database must be brought fully back online within 5 minutes
  • cChanges must be captured often enough that at most 5 minutes of data is unrecoverable✓
  • dEach backup file must be retained for at least 5 minutes before the system may delete it
Explanation:'At most 5 minutes of data lost' is an RPO. It constrains how often you capture changes — for example, shipping the transaction log every few minutes. Option (b) describes RTO (downtime), not data loss; (a) and (d) confuse RPO with backup duration and retention.

Topics

Backup Recovery 75Data Modeling 75Indexing 75Mongodb Aggregation 75Mongodb Indexes Queries 75Mongodb Operations 75Mongodb Replication Sharding 75Mongodb Schema Modeling 75Mongodb Transactions Consistency 75Mysql Indexes 75Mysql Innodb Transactions 75Mysql Operations 75Mysql Query Optimization 75Mysql Replication Scaling 75Mysql Types Constraints 75Nosql Models 75Performance Tuning 75Pg Indexes 75Pg Mvcc Vacuum 75Pg Operations 75Pg Query Planning 75Pg Transactions Locking 75Pg Types Constraints 75Query Optimization 75Redis Cluster Sharding 75Redis Data Structures 75Redis Expiration Eviction 75Redis Persistence 75Redis Replication Sentinel 75Redis Transactions Scripting 75Replication Scaling 75Security Access 75Transactions Isolation 75

By specialization

MongoDB 450MySQL 450PostgreSQL 450Redis 450

By seniority

Database JuniorDatabase MidDatabase Senior

Other fields

BackendFrontendFullstackDevOps / CloudAI EngineerMobileSecurityQA / Test AutomationData EngineerGame DevelopmentML Engineer

Test yourself against the 2475-question Database bank.

Start interview