Upload
arturo-donovan
View
214
Download
1
Embed Size (px)
Citation preview
1
Lecture 10: Transactions
2
The Setting
Database systems are normally being accessed by many users or processes at the same time. Both queries and modifications.
Unlike operating systems, which support interaction of processes, a DMBS needs to keep processes from troublesome interactions.
3
Example: Bad Interaction
You and your domestic partner each take $100 from different ATM’s at about the same time. The DBMS better make sure one
account deduction doesn’t get lost. Compare: An OS allows two people
to edit a document at the same time. If both write, one’s changes get lost.
4
ACID Transactions
A DBMS is expected to support “ACID transactions,” processes that are: Atomic : Either the whole process is done
or none is. Consistent : Database constraints are
preserved. Isolated : It appears to the user as if only
one process executes at a time. Durable : Effects of a process do not get
lost if the system crashes.
5
Transactions in SQL
SQL supports transactions, often behind the scenes. Each statement issued at the generic
query interface is a transaction by itself.
In programming interfaces like Embedded SQL or PSM, a transaction begins the first time a SQL statement is executed and ends with the program or an explicit transaction-end.
6
COMMIT
The SQL statement COMMIT causes a transaction to complete. It’s database modifications are now
permanent in the database.
7
ROLLBACK
The SQL statement ROLLBACK also causes the transaction to end, but by aborting. No effects on the database.
Failures like division by 0 or a constraint violation can also cause rollback, even if the programmer does not request it.
8
An Example: Interacting Processes
Assume the usual Sells(bar,beer,price) relation, and suppose that Joe’s Bar sells only Bud for $2.50 and Miller for $3.00.
Sally is querying Sells for the highest and lowest price Joe charges.
Joe decides to stop selling Bud and Miller, but to sell only Heineken at $3.50.
9
Sally’s Program
Sally executes the following two SQL statements, which we call (min) and (max), to help remember what they do.
(max)SELECT MAX(price) FROM SellsWHERE bar = ’Joe’’s Bar’;
(min) SELECT MIN(price) FROM SellsWHERE bar = ’Joe’’s Bar’;
10
Joe’s Program
At about the same time, Joe executes the following steps, which have the mnemonic names (del) and (ins).
(del) DELETE FROM Sells WHERE bar = ’Joe’’s Bar’;
(ins) INSERT INTO Sells VALUES(’Joe’’s Bar’, ’Heineken’,
3.50);
11
Interleaving of Statements
Although (max) must come before (min), and (del) must come before (ins), there are no other constraints on the order of these statements, unless we group Sally’s and/or Joe’s statements into transactions.
12
Example: Strange Interleaving
Suppose the steps execute in the order (max)(del)(ins)(min).
Joe’s Prices:Statement:Result:
Sally sees MAX < MIN!
2.50, 3.00
(del) (ins)
3.50
(min)
3.50
2.50, 3.00
(max)
3.00
13
Fixing the Problem by Using Transactions
If we group Sally’s statements (max)(min) into one transaction, then she cannot see this inconsistency.
She sees Joe’s prices at some fixed time. Either before or after he changes
prices, or in the middle, but the MAX and MIN are computed from the same prices.
14
Another Problem: Rollback
Suppose Joe executes (del)(ins), not as a transaction, but after executing these statements, thinks better of it and issues a ROLLBACK statement.
If Sally executes her statements after (ins) but before the rollback, she sees a value, 3.50, that never existed in the database.
15
Solution
If Joe executes (del)(ins) as a transaction, its effect cannot be seen by others until the transaction executes COMMIT. If the transaction executes ROLLBACK
instead, then its effects can never be seen.
16
Isolation Levels
SQL defines four isolation levels = choices about what interactions are allowed by transactions that execute at about the same time.
How a DBMS implements these isolation levels is highly complex, and a typical DBMS provides its own options.
17
Choosing the Isolation Level
Within a transaction, we can say:SET TRANSACTION ISOLATION LEVEL
Xwhere X =
1. SERIALIZABLE2. REPEATABLE READ3. READ COMMITTED4. READ UNCOMMITTED
18
Serializable Transactions
If Sally = (max)(min) and Joe = (del)(ins) are each transactions, and Sally runs with isolation level SERIALIZABLE, then she will see the database either before or after Joe runs, but not in the middle.
It’s up to the DBMS vendor to figure out how to do that, e.g.: True isolation in time. Keep Joe’s old prices around to answer
Sally’s queries.
19
Isolation Level Is Personal Choice
Your choice, e.g., run serializable, affects only how you see the database, not how others see it.
Example: If Joe Runs serializable, but Sally doesn’t, then Sally might see no prices for Joe’s Bar. i.e., it looks to Sally as if she ran in
the middle of Joe’s transaction.
20
Read-Commited Transactions
If Sally runs with isolation level READ COMMITTED, then she can see only committed data, but not necessarily the same data each time.
Example: Under READ COMMITTED, the interleaving (max)(del)(ins)(min) is allowed, as long as Joe commits. Sally sees MAX < MIN.
21
Repeatable-Read Transactions
Requirement is like read-committed, plus: if data is read again, then everything seen the first time will be seen the second time. But the second and subsequent reads
may see more tuples as well.
22
Example: Repeatable Read
Suppose Sally runs under REPEATABLE READ, and the order of execution is (max)(del)(ins)(min). (max) sees prices 2.50 and 3.00. (min) can see 3.50, but must also see
2.50 and 3.00, because they were seen on the earlier read by (max).
23
Read Uncommitted
A transaction running under READ UNCOMMITTED can see data in the database, even if it was written by a transaction that has not committed (and may never).
Example: If Sally runs under READ UNCOMMITTED, she could see a price 3.50 even if Joe later aborts.
24
Concurrency Control
T1 T2 … Tn
DB(consistencyconstraints)
25
Review
Why do we need transaction? What’s ACID? What’s SQL support for transaction? What’s the four isolation level
SERIALIZABLE REPEATABLE READ READ COMMITTED READ UNCOMMITTED
26
Example:
T1: Read(A) T2: Read(A)A A+100 A A2Write(A) Write(A)Read(B) Read(B)B B+100 B B2Write(B) Write(B)
Constraint: A=B
27
Schedule A
T1 T2Read(A); A A+100Write(A);Read(B); B B+100;Write(B);
Read(A);A A2;
Write(A);
Read(B);B B2;
Write(B);
A B25 25
125
125
250
250250 250
28
Schedule B
T1 T2
Read(A);A A2;
Write(A);
Read(B);B B2;
Write(B);Read(A); A A+100Write(A);Read(B); B B+100;Write(B);
A B25 25
50
50
150
150150 150
29
Schedule C
T1 T2Read(A); A A+100Write(A);
Read(A);A A2;
Write(A);Read(B); B B+100;Write(B);
Read(B);B B2;
Write(B);
A B25 25
125
250
125
250250 250
30
Schedule D
T1 T2Read(A); A A+100Write(A);
Read(A);A A2;
Write(A);
Read(B);B B2;
Write(B);Read(B); B B+100;Write(B);
A B25 25
125
250
50
150250 150
31
Schedule E
T1 T2’Read(A); A A+100Write(A);
Read(A);A A1;
Write(A);
Read(B);B B1;
Write(B);Read(B); B B+100;Write(B);
A B25 25
125
125
25
125125 125
Same as Schedule Dbut with new T2’
32
Want schedules that are “good”,regardless of
initial state and transaction semantics
Only look at order of read and writes
Example: Sc=r1(A)w1(A)r2(A)w2(A)r1(B)w1(B)r2(B)w
2(B)
33
Sc’=r1(A)w1(A) r1(B)w1(B)r2(A)w2(A)r2(B)w2(B)
T1 T2
Example: Sc=r1(A)w1(A)r2(A)w2(A)r1(B)w1(B)r2(B)w2(B)
34
Review
Why do we need transaction? What’s ACID? What’s SQL support for transaction? What’s the four isolation level
SERIALIZABLE REPEATABLE READ READ COMMITTED READ UNCOMMITTED
35
Example:
T1: Read(A) T2: Read(A)A A+100 A A2Write(A) Write(A)Read(B) Read(B)B B+100 B B2Write(B) Write(B)
Constraint: A=B
36
Schedule D
T1 T2Read(A); A A+100Write(A);
Read(A);A A2;
Write(A);
Read(B);B B2;
Write(B);Read(B); B B+100;Write(B);
A B25 25
125
250
50
150250 150
37
However, for Sd:Sd=r1(A)w1(A)r2(A)w2(A)
r2(B)w2(B)r1(B)w1(B)
as a matter of fact, T2 must precede T1
in any equivalent schedule, i.e., T2 T1
38
T1 T2 Sd cannot be rearrangedinto a serial
scheduleSd is not “equivalent” to
any serial scheduleSd is “bad”
T2 T1
Also, T1 T2
39
Schedule C
T1 T2Read(A); A A+100Write(A);
Read(A);A A2;
Write(A);Read(B); B B+100;Write(B);
Read(B);B B2;
Write(B);
A B25 25
125
250
125
250250 250
40
Returning to Sc
Sc=r1(A)w1(A)r2(A)w2(A)r1(B)w1(B)r2(B)w2(B)
T1 T2 T1 T2
no cycles Sc is “equivalent” to aserial schedule(in this case T1,T2)
41
Concepts
Transaction: sequence of ri(x), wi(x) actionsConflicting actions: r1(A) w2(A) w1(A)
w2(A) r1(A) w2(A)
Schedule: represents chronological orderin which actions are executed
Serial schedule: no interleaving of actions or transactions
42
Definition
S1, S2 are conflict equivalent schedulesif S1 can be transformed into S2 by a series of swaps on non-conflicting actions.
43
Definition
A schedule is conflict serializable if it is conflict equivalent to some serial schedule.
44
Nodes: transactions in SArcs: Ti Tj whenever
- pi(A), qj(A) are actions in S- pi(A) <S qj(A)
- at least one of pi, qj is a write
Precedence graph P(S) (S is
schedule)
45
Exercise:
What is P(S) forS = w3(A) w2(C) r1(A) w1(B) r1(C) w2(A) r4(A) w4(D)
Is S serializable?
46
Another Exercise:
What is P(S) forS = w1(A) r2(A) r3(A) w4(A) ?
47
Lemma
S1, S2 conflict equivalent P(S1)=P(S2)
Proof:Assume P(S1) P(S2) Ti: Ti Tj in S1 and not in S2
S1 = …pi(A)... qj(A)… pi, qj
S2 = …qj(A)…pi(A)... conflict
S1, S2 not conflict equivalent
48
Note: P(S1)=P(S2) S1, S2 conflict equivalent
Counter example:
S1=w1(A) r2(A) w2(B) r1(B) S2=r2(A) w1(A) r1(B) w2(B)
49
Theorem
P(S1) acyclic S1 conflict serializable
() Assume S1 is conflict serializable Ss: Ss, S1 conflict equivalent P(Ss) = P(S1)
P(S1) acyclic since P(Ss) is acyclic
50
() Assume P(S1) is acyclicTransform S1 as follows:(1) Take T1 to be transaction with no incident arcs(2) Move all T1 actions to the front
S1 = ……. qj(A)…….p1(A)…..
(3) we now have S1 = < T1 actions ><... rest ...>(4) repeat above steps to serialize rest!
T1
T2 T3
T4
TheoremP(S1) acyclic S1 conflict serializable