cs4432 recovery 1

CS4432: Database Systems IILecture #20

Failure Recovery

Professor Elke A. Rundensteiner

cs4432 recovery 2

Transaction: a collection of actions

that preserve consistency

Consistent DB Consistent DB’T

cs4432 recovery 3

Big assumption:

If T starts with consistent state ANDT executes in isolation

T leaves consistent state

cs4432 recovery 4

Our failure model

processor

memory disk

CPU

M D

cs4432 recovery 5

Operations reStorage Hierarchy:• Input (x): block containing x memory• Output (x): block containing x disk

• Read (x,t): do input(x) if necessary t value of x in block

• Write (x,t): do input(x) if necessary value of x in block t

Memory Disk

x x

cs4432 recovery 6

Key problem : Unfinished transaction

Example: Constraint: A=B T1: A A 2 B B 2

cs4432 recovery 7

T1: Read (A,t); t t2Write (A,t);Read (B,t); t t2Write (B,t);Output-to-Disk (A);Output-to-Disk (B);

A: 8B: 8

A: 8B: 8

memory disk

1616

16

failure!

cs4432 recovery 8

• Need atomicity: execute all actions of a transaction or none at all

cs4432 recovery 9

One solution: undo logging (immediate

modification)

A la Hansel and Gretel recording their navigation through forest via bread crumbs …

Must have durable undo logging !!!

cs4432 recovery 10

T1: Read (A,t); t t2 A=B

Write (A,t);Read (B,t); t t2Write (B,t);Output-to-disk (A);Output-to-disk (B);

A:8B:8

A:8B:8

memory disk log

Undo logging (Immediate modification)

1616

<T1, start><T1, A, 8>

<T1, commit>16 <T1, B, 8>

16

cs4432 recovery 11

One “complication”

• Log is first written in memory• Not written to disk on every action

memory

DB

Log

A: 8 16B: 8 16Log:<T1,start><T1, A, 8><T1, B, 8>

A: 8B: 8

16BAD STATE

# 1

cs4432 recovery 12

One “complication”

• Log is first written in memory• Not written to disk on every action

memory

DB

Log

A: 8 16B: 8 16Log:<T1,start><T1, A, 8><T1, B, 8><T1, commit>

A: 8B: 8

16BAD STATE

# 2

<T1, B, 8><T1, commit>

...

cs4432 recovery 13

Undo logging rules

(1)For every action generate undo log record (containing old value)

(2) Before x is modified on disk, log records pertaining to x must be

on disk (write ahead logging)(3) Before commit is written to log on

disk, all writes of transaction must be

reflected on disk

cs4432 recovery 14

Recovery rules: Undo logging• For every Ti with <Ti, start> in log:

- If <Ti,commit> or <Ti,abort> in log, do nothing

- Else For all <Ti, X, v> in log:write (X, v)output (X )

Write <Ti, abort> to log

IS THIS CORRECT??

cs4432 recovery 15

Recovery rules: Undo logging

(1) Let S = set of transactions with<Ti, start> in log, but no<Ti, commit> (or <Ti, abort>) record

in log(2) For each <Ti, X, v> in log,

in reverse order (latest earliest) do: - write old value from log back to disk:

- if Ti S then - write (X, v) - output (X)

(3) For each Ti S do- write <Ti, abort> to log

cs4432 recovery 16

What if failure during recovery?No problem! Undo idempotent

cs4432 recovery 17

To discuss next :

• Redo logging• Undo/redo logging, why both?• Checkpoints

cs4432 recovery 18

Redo logging (deferred modification)

T1: Read(A,t); t t2; write (A,t); Read(B,t); t t2; write (B,t);

Output(A); Output(B)

A: 8B: 8

A: 8B: 8

memory DB LOG

1616

<T1, start><T1, A, 16><T1, B, 16>

<T1, commit>

output16

cs4432 recovery 19

Redo logging rules

(1) For every action, generate redo logrecord (containing new value)

(2) Before X is modified on disk (DB),all log records for transaction that

modified X (including commit) must be on disk

(3) Flush log at commit

cs4432 recovery 20

• Idea: Need to redo if transaction commit is in log because we don’t know if transaction failed or not.

• For every Ti with <Ti, commit> in log:– For all <Ti, X, v> in log:

Write(X, v)Output(X)

Recovery rules: Redo logging

IS THIS CORRECT??

cs4432 recovery 21

(1) Let S = set of transactions with<Ti, commit> in log

(2) For each <Ti, X, v> in log, in forward order (earliest latest) do:

- if Ti S then Write(X, v) Output(X)

optional

Recovery rules: Redo logging

cs4432 recovery 22

Recovery is very, very

SLOW !Redo log:

First T1 wrote A,B LastRecord Committed a year ago

Record(1 year ago) --> STILL, Need to redo after crash!!

... ... ...

Crash

cs4432 recovery 23

Solution: Checkpoint (simple version)

Periodically:(1) Do not accept new transactions(2) Wait until all transactions finish(3) Flush all log records to disk (log)(4) Flush all buffers to disk (DB) (do not discard buffers)

(5) Write “checkpoint” record on disk (log)

(6) Resume transaction processing

cs4432 recovery 24

Example: what to do at recovery?

Redo log (disk):

<T1

,A,1

6>

<T1

,com

mit

>

Ch

eck

poin

t

<T2

,B,1

7>

<T2

,com

mit

>

<T3

,C,2

1>

Crash... ... ... ...

...

...

cs4432 recovery 25

Key drawbacks:

• Undo logging: cannot bring backup DBcopies up to date

• Redo logging: need to keep all modified blocks inmemory

until commit

cs4432 recovery 26

Solution: undo/redo logging!

Combine Undo/Redo Logging, namely:

Update <Ti, Xid, New X val, Old X val>page X

cs4432 recovery 27

Rules

• Page X can be flushed before orafter Ti commit

• Log record flushed before corresponding updated page (WAL)

• Flush at commit (log only)

cs4432 recovery 28

Non-quiesce checkpoint

LOG

for undo dirty buffer

pool pagesflushed

Start-ckptactive TR:

Ti,T2,...

endckpt

.........

...

cs4432 recovery 29

Examples what to do at recovery time?

no T1 commit

LOG

T1,-a

...CkptT1

...Ckptend

...T1-b

...

Undo T1 (undo a,b)

cs4432 recovery 30

Example

LOG

...T1

a... ...

T1

b... ...

T1

c...

T1

cmt...

ckpt-end

ckpt-sT1

Redo T1: (redo b,c)

cs4432 recovery 31

Recovery process:• Backwards pass (end of log latest checkpoint

start)

– construct set S of committed transactions– undo actions of transactions not in S

• Undo pending transactions– follow undo chains for transactions in

(checkpoint active list) - S

• Forward pass (latest checkpoint start end of log)

– redo actions of S transactions

backward pass

forward passstart

check-point

cs4432 recovery 32

Summary

• Consistency of data• One source of problems: failures

- Logging- Redundancy

• Another source of problems: Data Sharing..... next