Recovery Management

Q1. Database recovery is the process of:

1. A.Deleting all data from the tables in the entire database
2. B.Restoring the database to a correct state after a failure✓ Correct
3. C.Creating new tables and defining their column structures
4. D.Optimizing queries by choosing better execution plans

Report an error in this question

Q2. A transaction log (write-ahead log) records:

1. A.Table structures and their column definitions in the database schema
2. B.Only SELECT queries that have been executed against the database tables
3. C.User login information and authentication credentials stored securely
4. D.All changes made by transactions before they are applied to the database✓ Correct

Report an error in this question

Q3. The redo operation:

1. A.Creates a new database from scratch without using any data from the previous state
2. B.Re-applies changes of committed transactions that may not have been written to disk✓ Correct
3. C.Undoes all changes of committed transactions and reverts to the original data state
4. D.Deletes the entire log file and removes all records of previous transaction actions

Report an error in this question

Q4. The undo operation:

1. A.Reverses changes made by uncommitted transactions✓ Correct
2. B.Creates a backup of the current database on disk
3. C.Optimizes queries for faster execution performance
4. D.Applies committed changes to the database on disk

Report an error in this question

Q5. A checkpoint in database recovery is:

1. A.A type of lock used to prevent concurrent access to shared data items
2. B.An SQL statement used to query data from one or more database tables
3. C.A point at which all modified data is written to disk and recorded in the log✓ Correct
4. D.A type of index structure built on columns for faster data retrieval

Report an error in this question

Q6. The Write-Ahead Log (WAL) protocol requires:

1. A.Logs are only written to disk after the transaction has committed fully
2. B.No log is needed for any database operations under this recovery protocol
3. C.Data must be written to disk before the corresponding log records are made
4. D.Log records must be written to disk before the corresponding data changes✓ Correct

Report an error in this question

Q7. A system crash typically causes:

1. A.Loss of data in main memory but not on disk✓ Correct
2. B.No data loss of any kind after the crash
3. C.Loss of all data everywhere in the system
4. D.Loss of data on disk but not in main memory

Report an error in this question

Q8. A media failure refers to:

1. A.Application crash caused by a software defect
2. B.Network slowness affecting query response times
3. C.Disk or storage device failure causing data loss✓ Correct
4. D.Memory overflow due to excessive data loading

Report an error in this question

Q9. Database backups are important for:

1. A.Reducing table size by removing unnecessary data rows
2. B.Recovering from media failures and catastrophic events✓ Correct
3. C.Creating new databases from the backed up data copies
4. D.Improving query speed and reducing execution latency

Report an error in this question

Q10. An incremental backup stores:

1. A.The entire database every single time
2. B.Only the database schema definitions
3. C.Only the changes since the last backup✓ Correct
4. D.Only the transaction log file entries

Report an error in this question

Q11. Deferred update (no-undo/redo) recovery technique:

1. A.Requires neither undo nor redo operations during the recovery process
2. B.Writes changes immediately and may need undo for uncommitted changes
3. C.Does not write changes to disk until commit, so undo is never needed✓ Correct
4. D.Does not use a log file for recording any of the transaction changes

Report an error in this question

Q12. Immediate update recovery technique:

1. A.Does not use a log file for any of the transaction record entries
2. B.Writes changes only after commit, requiring only redo operations
3. C.Requires only redo operations during the crash recovery process
4. D.Writes changes to disk before commit, requiring both undo and redo✓ Correct

Report an error in this question

Q13. Shadow paging recovery technique:

1. A.Maintains two page tables: current and shadow✓ Correct
2. B.Requires checkpoints at all times to function
3. C.Uses only a log file for the recovery process
4. D.Cannot handle any system crashes that may occur

Report an error in this question

Q14. After a crash, which transactions need to be redone?

1. A.All transactions regardless of their commit state now
2. B.Committed transactions whose changes may not be on disk✓ Correct
3. C.Uncommitted transactions that were active during crash
4. D.Only aborted transactions that failed before the crash

Report an error in this question

Q15. After a crash, which transactions need to be undone?

1. A.All transactions regardless of their state at the time of the crash
2. B.No transactions need to be undone during the recovery after a crash
3. C.Transactions that were active (not committed) at the time of crash✓ Correct
4. D.Committed transactions that completed successfully before the crash

Report an error in this question

Q16. The ARIES recovery algorithm stands for:

1. A.Automated Recovery and Integrity Enforcement System
2. B.Advanced Recovery and Index Enhancement System
3. C.Algorithm for Redo and Integrity Evaluation System
4. D.Algorithm for Recovery and Isolation Exploiting Semantics✓ Correct

Report an error in this question

Q17. ARIES uses which three phases during recovery?

1. A.Scan, Parse, Execute
2. B.Analysis, Redo, Undo✓ Correct
3. C.Lock, Execute, Unlock
4. D.Read, Write, Commit

Report an error in this question

Q18. A log sequence number (LSN) is:

1. A.The total number of active transactions
2. B.A type of lock used in concurrency control
3. C.The size of the log file on stable storage
4. D.A unique identifier for each log record✓ Correct

Report an error in this question

Q19. Fuzzy checkpointing differs from standard checkpointing in that:

1. A.No log entry is written to record the occurrence of the checkpoint
2. B.The database does not need to stop processing during the checkpoint✓ Correct
3. C.All processing must stop completely during the checkpoint operation
4. D.It occurs only once during the entire lifetime of the database

Report an error in this question

Q20. Physiological logging means:

1. A.Log records describe changes at the page level with logical operations within the page✓ Correct
2. B.Only logical operation names are logged without any reference to the physical pages
3. C.Only physical byte-level changes are logged for each modification to the data pages
4. D.No logging occurs at all for any modification made to the database during processing

Report an error in this question

Q21. In ARIES, the analysis phase determines:

1. A.The set of dirty pages and active transactions at crash time✓ Correct
2. B.Only committed transactions that completed before the crash
3. C.The hardware failure cause and the affected disk components
4. D.Only the last checkpoint that was recorded in the log file

Report an error in this question

Q22. In ARIES, the redo phase:

1. A.Replays all logged changes from the earliest dirty page LSN forward✓ Correct
2. B.Starts from the end of the log and processes records in reverse
3. C.Only replays committed transactions and ignores uncommitted ones
4. D.Undoes all transactions that were active at the time of the crash

Report an error in this question

Q23. In ARIES, Compensation Log Records (CLRs) are:

1. A.Checkpoint records that mark points of database consistency
2. B.Log records for committed transactions only in the system
3. C.Log records written during undo to record the undo actions✓ Correct
4. D.Index maintenance records for B-tree structure updates

Report an error in this question

Q24. The steal/no-force buffer management policy means:

1. A.Uncommitted pages can be written to disk (steal); committed need not be forced immediately (no-force)✓ Correct
2. B.Committed pages are always forced to disk immediately; uncommitted pages are never written to disk
3. C.Only committed pages are present on disk; uncommitted data remains exclusively in the memory buffer
4. D.All pages are always kept in memory; nothing is written to disk until the entire database shuts down

Report an error in this question

Q25. No-steal/force policy would mean:

1. A.Uncommitted pages are never written to disk, and committed pages are always forced to disk✓ Correct
2. B.Neither undo nor redo is needed with this policy since both cases are eliminated
3. C.Uncommitted pages can be written to disk at any time before the transaction commits
4. D.Committed pages may not be on disk immediately and can remain only in memory buffer

Report an error in this question

Q26. The redoLSN in ARIES is:

1. A.The checkpoint LSN that marks the most recent consistency checkpoint
2. B.The last LSN that was recorded in the transaction log before the crash
3. C.The LSN from which redo must start (earliest recLSN among dirty pages)✓ Correct
4. D.The LSN of the last commit record written to the stable storage log

Report an error in this question

Q27. Write-ahead logging requires that log records be written to stable storage before:

1. A.The transaction starts its initial read operation
2. B.Any query is executed against the database tables
3. C.The corresponding data page is flushed to disk✓ Correct
4. D.The user submits the query to the database system

Report an error in this question

Q28. In ARIES, when a transaction is undone, the undo follows:

1. A.The commit order of other transactions in the same schedule
2. B.The table creation order defined in the database schema file
3. C.The lastLSN chain backward through that transaction's log records✓ Correct
4. D.A random order without any specific sequence or priority rules

Report an error in this question

Q29. Remote backup systems for disaster recovery provide:

1. A.A copy of the database at a geographically separate site✓ Correct
2. B.Better index structures for improving search performance
3. C.Improved normalization of tables to reduce data redundancy
4. D.Faster query processing through parallel execution plans

Report an error in this question

Q30. The concept of idempotent redo means:

1. A.Applying redo multiple times produces the same result as applying it once✓ Correct
2. B.Redo can only be applied once and cannot be repeated after that point
3. C.Redo changes the result each time it is applied to the database state
4. D.Redo is never repeated because recovery always succeeds on first try

Report an error in this question