Chapter 3

Database Recovery Techniques

 Chapter 19 Outline
Databases Recovery
1. Purpose of Database Recovery
2. Types of Failure
3 . Transaction Log
4 . Data Updates
5. Data Caching
6 . Transaction Roll-back (Undo) and Roll-Forward
7 . Checkpointing
8 . Recovery schemes
9 . ARIES Recovery Scheme
10. Recovery in Multi database System

Slide 19- 2
 Database Recovery
1 Purpose of Database Recovery
 To bring the database into the last consistent state, which existed prior to
the failure.
 To preserve transaction properties (Atomicity, Consistency, Isolation and
Durability).
 Example:
 If the system crashes before a fund transfer transaction completes its
execution, then either one or both accounts may have incorrect value.
Thus, the database must be restored to the state before the transaction
modified any of the accounts.

Slide 19- 3
Database Recovery
2 Types of Failure
 The database may become unavailable for use due to
 Transaction failure: Transactions may fail because of
incorrect input, deadlock, incorrect synchronization.
 System failure: System may fail because of addressing
error, application error, operating system fault, RAM
failure, etc.
 Media failure: Disk head crash, power disruption, etc.
Slide 19- 4
Database Recovery
3 Transaction Log
 For recovery from any type of failure data values prior to
modification (BFIM - BeFore Image) and the new value after
modification (AFIM – AFter Image) are required.
 These values and other information is stored in a sequential file
called Transaction log. A sample log is given below. Back P and
Next P point to the previous and next log records of the same
transaction.
T ID Back P Next P Operation Data item BFIM AFIM
T1 0 1 Begin
T1 1 4 Write X X = 100 X = 200
T2 0 8 Begin
T1 2 5 W Y Y = 50 Y = 100
T1 4 7 R M M = 200 M = 200
T3 0 9 R N N = 400 N = 400
T1 5 nil End

Slide 19- 5
Database Recovery
4 Data Update
 Immediate Update: As soon as a data item is modified in cache, the disk copy is
updated.
 Deferred Update: All modified data items in the cache is written either after a
transaction ends its execution or after a fixed number of transactions have
completed their execution.
 Shadow update: The modified version of a data item does not overwrite its disk
copy but is written at a separate disk location.
 In-place update: The disk version of the data item is overwritten by the cache
version.

Slide 19- 6
Database Recovery
5 Data Caching
 Data items to be modified are first stored into database
cache by the Cache Manager (CM) and after modification
they are flushed (written) to the disk.
 The flushing is controlled by Modified and Pin-Unpin bits.
 Pin-Unpin: Instructs the operating system not to flush the data item.
 Modified: Indicates the AFIM of the data item.

Slide 19- 7
 Database Recovery
6 Transaction Roll-back (Undo) and Roll-Forward
(Redo)
 To maintain atomicity, a transaction’s operations are
redone or undone.
 Undo: Restore all BFIMs on to disk (Remove all
AFIMs).
 Redo: Restore all AFIMs on to disk.
 Database recovery is achieved either by performing
only Undos or only Redos or by a combination of the
two. These operations are recorded in the log as they
happen.

Slide 19- 8
Database Recovery

Slide 19- 9
Database Recovery

Slide 19- 10
Database Recovery
Roll-back: One execution of T1, T2 and T3 as recorded in
the log.

Slide 19- 11
Database Recovery
Write-Ahead Logging
 When in-place update (immediate or deferred) is used then log is necessary
for recovery and it must be available to recovery manager. This is achieved
by Write-Ahead Logging (WAL) protocol. WAL states that
 For Undo: Before a data item’s AFIM is flushed to the database disk
(overwriting the BFIM) its BFIM must be written to the log and the log
must be saved on a stable store (log disk).
 For Redo: Before a transaction executes its commit operation, all its
AFIMs must be written to the log and the log must be saved on a stable
store.

Slide 19- 12
Database Recovery
7 Checkpointing
 Time to time (randomly or under some criteria) the database flushes its buffer
to database disk to minimize the task of recovery. The following steps defines
a checkpoint operation:
1. Suspend execution of transactions temporarily.

2. Force write modified buffer data to disk.

3. Write a [checkpoint] record to the log, save the log to disk.

4. Resume normal transaction execution.

 During recovery redo or undo is required to transactions appearing after
[checkpoint] record.

Slide 19- 13
Database Recovery
Steal/No-Steal and Force/No-Force
 Possible ways for flushing database cache to database disk:
1. Steal: Cache can be flushed before transaction commits.

2. No-Steal: Cache cannot be flushed before transaction commit.

3. Force: Cache is immediately flushed (forced) to disk.

4. No-Force: Cache is deferred until transaction commits

 These give rise to four different ways for handling recovery:

 Steal/No-Force (Undo/Redo)
 Steal/Force (Undo/No-redo)
 No-Steal/No-Force (Redo/No-undo)
 No-Steal/Force (No-undo/No-redo)

Slide 19- 14
Database Recovery
8 Recovery Scheme
 Deferred Update (No Undo/Redo)

 The data update goes as follows:

 A set of transactions records their updates in the log.
 At commit point under WAL scheme these updates are
saved on database disk.
 After reboot from a failure the log is used to redo all the
transactions affected by this failure. No undo is
required because no AFIM is flushed to the disk before
a transaction commits.

Slide 19- 15
Database Recovery
 Deferred Update in a single-user system
There is no concurrent data sharing in a single user system. The
data update goes as follows:
 A set of transactions records their updates in the log.
 At commit point under WAL scheme these updates are saved on
database disk.
 After reboot from a failure the log is used to redo all the
transactions affected by this failure. No undo is required
because no AFIM is flushed to the disk before a transaction
commits.

Slide 19- 16
Database Recovery

Slide 19- 17
Database Recovery
Deferred Update with concurrent users
 This environment requires some concurrency control

mechanism to guarantee isolation property of

transactions. In a system recovery transactions which
were recorded in the log after the last checkpoint were
redone. The recovery manager may scan some of the
transactions recorded before the checkpoint to get the
AFIMs.

Slide 19- 18
Database Recovery

Slide 19- 19
Database Recovery
Deferred Update with concurrent users
 Two tables are required for implementing this protocol:
 Active table: All active transactions are entered in this
table.
 Commit table: Transactions to be committed are entered in
this table.

 During recovery, all transactions of the commit table are

redone and all transactions of active tables are ignored
since none of their AFIMs reached the database. It is
possible that a commit table transaction may be redone
twice but this does not create any inconsistency because
of a redone is “idempotent”, that is, one redone for an
AFIM is equivalent to multiple redone for the same AFIM.

Slide 19- 20
Database Recovery
Recovery Techniques Based on Immediate Update
 Undo/No-redo Algorithm

 In this algorithm AFIMs of a transaction are flushed

to the database disk under WAL before it commits.
 For this reason the recovery manager undoes all
transactions during recovery.
 No transaction is redone.
 It is possible that a transaction might have
completed execution and ready to commit but this
transaction is also undone.

Slide 19- 21
Database Recovery
Recovery Techniques Based on Immediate Update
 Undo/Redo Algorithm (Single-user environment)

 Recovery schemes of this category apply undo and

also redo for recovery.
 In a single-user environment no concurrency control
is required but a log is maintained under WAL.
 Note that at any time there will be one transaction in
the system and it will be either in the commit table
or in the active table.
 The recovery manager performs:
 Undo of a transaction if it is in the active table.
 Redo of a transaction if it is in the commit table.

Slide 19- 22
Database Recovery
Recovery Techniques Based on Immediate Update
 Undo/Redo Algorithm (Concurrent execution)

 Recovery schemes of this category applies undo and also

redo to recover the database from failure.

 In concurrent execution environment a concurrency
control is required and log is maintained under WAL.
 Commit table records transactions to be committed and

active table records active transactions. To minimize the

work of the recovery manager checkpointing is used.
 The recovery performs:

 Undo of a transaction if it is in the active table.

 Redo of a transaction if it is in the commit table.

Slide 19- 23
Database Recovery
Shadow Paging
 The AFIM does not overwrite its BFIM but recorded at

another place on the disk. Thus, at any time a data item

has AFIM and BFIM (Shadow copy of the data item) at
two different places on the disk.

X Y
X' Y'

Database

X and Y: Shadow copies of data items

X' and Y': Current copies of data items
Slide 19- 24
Database Recovery
Shadow Paging
 To manage access of data items by concurrent
transactions two directories (current and shadow) are
used.
 The directory arrangement is illustrated below. Here a page
is a data item.

Slide 19- 25
Database Recovery
The ARIES Recovery Algorithm
 The ARIES Recovery Algorithm is based on:

 WAL (Write Ahead Logging)

 Repeating history during redo:
 ARIES will retrace all actions of the database
system prior to the crash to reconstruct the
database state when the crash occurred.
 Logging changes during undo:
 It will prevent ARIES from repeating the completed
undo operations if a failure occurs during recovery,
which causes a restart of the recovery process.

Slide 19- 26
Database Recovery
The ARIES Recovery Algorithm (contd.)
 The ARIES recovery algorithm consists of three steps:
1. Analysis: step identifies the dirty (updated) pages in the
buffer and the set of transactions active at the time of
crash. The appropriate point in the log where redo is to
start is also determined.
2. Redo: necessary redo operations are applied.
3. Undo: log is scanned backwards and the operations of
transactions active at the time of crash are undone in
reverse order.

Slide 19- 27
Database Recovery
The ARIES Recovery Algorithm (contd.)
 The Log and Log Sequence Number (LSN)

 A log record is written for:

 (a) data update
 (b) transaction commit
 (c) transaction abort
 (d) undo
 (e) transaction end
 In the case of undo a compensating log record is
written.

Slide 19- 28
Database Recovery
The ARIES Recovery Algorithm (contd.)
 The Log and Log Sequence Number (LSN) (contd.)

 A unique LSN is associated with every log record.

 LSN increases monotonically and indicates the disk address
of the log record it is associated with.
 In addition, each data page stores the LSN of the latest log
record corresponding to a change for that page.
 A log record stores
 (a) the previous LSN of that transaction
 (b) the transaction ID
 (c) the type of log record.

Slide 19- 29
Database Recovery
The ARIES Recovery Algorithm (contd.)
 The Log and Log Sequence Number (LSN) (contd.)
 A log record stores:
1. Previous LSN of that transaction: It links the log record of each
transaction. It is like a back pointer points to the previous record
of the same transaction
2. Transaction ID
3. Type of log record
 For a write operation the following additional information is logged:
1. Page ID for the page that includes the item
2. Length of the updated item
3. Its offset from the beginning of the page
4. BFIM of the item
5. AFIM of the item

Slide 19- 30
Database Recovery
The ARIES Recovery Algorithm (contd.)
 The Transaction table and the Dirty Page table

 For efficient recovery following tables are also

stored in the log during checkpointing:
 Transaction table: Contains an entry for each
active transaction, with information such as
transaction ID, transaction status and the LSN of
the most recent log record for the transaction.
 Dirty Page table: Contains an entry for each dirty
page in the buffer, which includes the page ID and
the LSN corresponding to the earliest update to that
page.

Slide 19- 31
Database Recovery
The ARIES Recovery Algorithm (contd.)
 Checkpointing

 A checkpointing does the following:

 Writes a begin_checkpoint record in the log
 Writes an end_checkpoint record in the log. With this record
the contents of transaction table and dirty page table are
appended to the end of the log.
 Writes the LSN of the begin_checkpoint record to a special
file. This special file is accessed during recovery to locate the
last checkpoint information.
 To reduce the cost of checkpointing and allow the system to
continue to execute transactions, ARIES uses “fuzzy
checkpointing”.

Slide 19- 32
Database Recovery
The ARIES Recovery Algorithm (contd.)
 The following steps are performed for recovery
 Analysis phase: Start at the begin_checkpoint record and

proceed to the end_checkpoint record. Access transaction table

and dirty page table are appended to the end of the log. Note that
during this phase some other log records may be written to the log
and transaction table may be modified. The analysis phase
compiles the set of redo and undo to be performed and ends.
 Redo phase: Starts from the point in the log up to where all dirty

pages have been flushed, and move forward to the end of the log.
Any change that appears in the dirty page table is redone.
 Undo phase: Starts from the end of the log and proceeds

backward while performing appropriate undo. For each undo it

writes a compensating record in the log.
 The recovery completes at the end of undo phase.

Slide 19- 33
Database Recovery

Slide 19- 34
Database Recovery
10 Recovery in multidatabase system
 A multidatabase system is a special distributed database system
where one node may be running relational database system under
UNIX, another may be running object-oriented system under
Windows and so on.
 A transaction may run in a distributed fashion at multiple nodes.
 In this execution scenario the transaction commits only when all
these multiple nodes agree to commit individually the part of the
transaction they were executing.
 This commit scheme is referred to as “two-phase commit” (2PC).
 If any one of these nodes fails or cannot commit the part of the

transaction, then the transaction is aborted.

 Each node recovers the transaction under its own recovery protocol.

Slide 19- 35
Summary
 Databases Recovery
 Types of Failure
 Transaction Log
 Data Updates
 Data Caching
 Transaction Roll-back (Undo) and Roll-Forward
 Checkpointing
 Recovery schemes
 ARIES Recovery Scheme
 Recovery in Multidatabase System

Slide 19- 36