Lecture 12: Concurrency Control,

Database Recovery

1
Logistics
 Final report: Due on 12/15/24
 Extra points
 Teaching evaluation screenshot (ONLY SHOW
SUBMISSION and NO SCORE)

 Final exam
 12/16/24 online _ 6 to 8 pm

2
Topics
 Transaction processing
 Concurrency control
 Database recovery

3
Why Need Understanding
Transaction Processing
 Multi-user database
 More than one user processing the database at
the same time
 Controlling read and update order
 Issues
 How can we prevent users from interfering with
each other’s work ?
 How can we safely process transactions on the
database without corrupting or losing data ?
 If there is a problem (e.g., power failure or system
crash), how can we recover without losing all of
our data ?

4
Transaction Processing
 A transaction is a set
of read and write operations that must
either commit or abort.
 Commit point (Success)
 all actions are permanently saved in the database
 Not commit: dirty
 Abort (Failure)
 none of the actions are saved
 Such a transaction is rolled back or undone.

5
ACID Properties of Transactions
: relational database VS NoSQL
database
 Atomicity ‘All or nothing’ property. Commit or
abort
 Consistency Must transform database from one
consistent state to another.
 Isolation Partial effects of incomplete
transactions should not be visible to other
transactions.
 Durability Effects of a committed transaction
are permanent and must not be lost because of
later failure.
 Dirty read problem: isolation
 Recovery tries to satisfy which property?
6
Example
 Two users executing similar transactions
 Case 1:
User A User B
Read Salary for emp 101 Read Salary for emp 101
Multiply salary by 1.03 Multiply salary by 1.04
Write Salary for emp Write Salary for emp
101 101

 Case2:
User A User B
Read inventory for Prod Read inventory for Prod
P200 P200
Decrement inventory by Decrement inventory by
5 7
7 Write inventory for Prod Write inventory for Prod
P200 P200
Four General classes of problems
with Transaction Processing
 The Lost Update Problem
 Successfully completed update is overridden by
another user.
 The Dirty Read Problem (also called uncommitted
dependency)
 Occurs when one transaction can see intermediate
results of another transaction before this
transaction has committed.
 The Incorrect Analysis Problem
 Occurs when transaction reads several values but
second transaction updates some of them during
execution of first.
 The Non-Repeatable Read Problem
 nonrepeatable read occurs when a transaction reads the
8 same row twice but gets different data each time
Serial Schedules and
Serializability
 Can we insist only one transact at a
time?
 Transaction throughput: The number of
transactions we can perform in a given time
period. Often reported as Transactions per
second or TPS.
 Serializable
 More concurrent transactions = serial order (one
after another).
 Schedule: Sequence of reads/writes by set of
concurrent transactions.

9 Serial Schedule: Schedule where operations of

10
Concurrency Control and Locking
 Concurrency Control is a method for
controlling or scheduling the operations in
such a way that concurrent transactions can
be executed.
 Concurrent transactions need to be serialized
 Locking
 Locking is done to data items in order to reserve
them for future operations.
 A lock is a logical flag set by a transaction to alert
other transactions the data item is in use.
 Timestamping.

11
Characteristics of Locks
 Two types:
 Implicit Locks are applied by the DBMS
Explicit Locks are applied by application
programs.
 Lock granularity
 a single data item (value)
 an entire row of a table
 a page (memory segment) (many rows worth)
 an entire table or database
 Requirements of transaction
 An Exclusive Lock (XL) prevents any other
transaction from reading or modifying the locked
item.
12
 A Shared Lock (SL) allows another transaction to
Locking - Basic Rules
 If transaction has shared lock on item, can
read but not update item.
 If transaction has exclusive lock on item, can
both read and update item.
 Reads cannot conflict, so more than one
transaction can hold shared locks
simultaneously on same item.
 Exclusive lock gives transaction exclusive
access to that item.

13
Two phase of locking
 2PL has two phases:
 A transaction acquires locks on data items it will
need to complete the transaction. This is called
the growing phase.
 Once one lock is released, all no other lock may be
acquired. This is called the shrinking phase.

14
 Amy and Bill are two employees who will get a 5% raise this year.
Carl will get a 3% raise this year.

15
Deadlock
 Updating two different data items (P200 and
P300).

16
Solution of Dead Lock
 Two main ways to deal with deadlock.
 Prevent it in the first place by giving each
transaction exclusive rights to acquire all locks
needed before proceeding.
 Allow the deadlock to occur, then break it by
aborting one of the transactions.

17
Timestamping
 Ordering transaction globally
 Smaller timestamps, get priority in the event
of conflict.
 Conflict is resolved by rolling back and
restarting transaction.
 No locks so no deadlock.

18
Timestamping
Timestamp
A unique identifier created by DBMS that
indicates relative starting time of a transaction.
 Can be generated by using system clock at
time transaction started, or by incrementing
a logical counter every time a new
transaction starts.

19
Timestamping
 Read/write proceeds only if last update on
that data item was carried out by an older
transaction.
 Otherwise, transaction requesting read/write
is restarted and given a new timestamp.
 Also timestamps for data items:
 read-timestamp - timestamp of last transaction to
read item;
 write-timestamp - timestamp of last transaction
to write item.

20
Transaction Interruption
 There are many situations in which a
transaction may not reach a commit or abort
point.
 An operating system crash
 The DBMS can crash
 The system might lose power
 A disk may fail or other hardware may fail.
 Human error can result in deletion of critical data.

21
Database Recovery
 The process of restoring the database and the
data to a consistent state.
 This may include restoring lost data up to the
point of the event (e.g. system crash).
 Which ACID state?
 Durablity: committed (saved) won’t be lost
 Two types of storage
 volatile (main memory) and nonvolatile.

22
Recovery Facilities
 DBMS should provide following facilities to assist
with recovery:
 Backup mechanism
 Backup mechanism, which makes periodic backup
copies of database.
 Logging facilities
 Logging facilities, which keep track of current state of
transactions and database changes.
 Checkpoint facility
 Checkpoint facility, which enables updates to
database in progress to be made permanent.
 Recovery manager, which allows DBMS to
restore database to consistent state following
a failure.
Pearson Education © 2014 23
Recovering approach (Manual
Reprocessing)
 Manual Reprocessing
 Database is periodically backed up (a
database save) and all transactions applied since
the last save are recorded
 When system crashes
 the latest database backup set is restored to the point
just before the crash.
 Several shortcomings to the Manual
Reprocessing approach:
 Time required to re-apply transactions
 Transactions might have other (physical)
consequences
 Re-applying concurrent transactions in the same
24 original sequence may not be possible.
Recovering approach (Automated
Recovery)
 Log file (journal)
 a file separate from the data that records all of the
changes made to the database by transactions.
 Contains information about all updates to
database:
 Transaction records.
 Checkpoint records.
 Often used for other purposes (for example,
auditing).
 Log information
 Before Image: A copy of the table record (or data
item) before it was changed by the transaction.
 After Image: A copy of the table record (or data
25 item) after it was changed by the transaction.
Sample Log File

26
Automated Recovery process
 Operations
 Rollback: Undo any partially completed
transactions (ones in progress when the crash
occurred) by applying the before images to the
database.
 Rollforward: Redo the transactions by applying
the after images to the database. This is done for
transactions that were committed before the
crash.
 Automated Recovery process
 rollback and rollforward to restore the database

27
Checkpointing
 Checkpoints
 Point of synchronization between database and log
file.
 between database saves
 The DBMS flushes all pending transactions and
writes all data to disk and transaction log.
 Database can be recovered from the last
checkpoint in much less time.

28
Example
 Physical backup of the data just before
Transaction A begins
 Amy = 45
 Bill = 38
 Carl = 51
 R_R = .05

29
Transaction A Begin
Transaction B Begin
Transaction C Begin
Transaction B Write: Bill Before: 38 After: 39.9
Transaction A Write: Amy Before: 45 After: 47.7
Transaction B Commit
Transaction C Write: R_R Before .05 After: .03

30