This document discusses database systems and transactions. It contains 3 key points:
1. Transactions allow concurrent execution of processes by interleaving their execution and switching the CPU between processes during I/O operations, preventing any single long process from delaying others.
2. A transaction includes one or more database operations like insert, delete, retrieve, etc. and must preserve consistency when executing concurrently with other transactions.
3. Scheduling of transactions and concurrency control techniques like locking are used to regulate interactions between concurrently executing transactions and ensure consistency is preserved. Transactions are executed in buffers in main memory and their operations are written to a log on disk.
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0 ratings0% found this document useful (0 votes)
19 views66 pages
Transaction Concurrency
This document discusses database systems and transactions. It contains 3 key points:
1. Transactions allow concurrent execution of processes by interleaving their execution and switching the CPU between processes during I/O operations, preventing any single long process from delaying others.
2. A transaction includes one or more database operations like insert, delete, retrieve, etc. and must preserve consistency when executing concurrently with other transactions.
3. Scheduling of transactions and concurrency control techniques like locking are used to regulate interactions between concurrently executing transactions and ensure consistency is preserved. Transactions are executed in buffers in main memory and their operations are written to a log on disk.
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 66
TEDU – CMPE 232 – Database Systems
• A and B, executing concurrently in an interleaved fashion.
• Interleaving keeps the CPU busy when a process requires an input or output (I/O) operation, such as reading a block from disk. • The CPU is switched to execute another process rather than remaining idle during I/O time. • Interleaving also prevents a long process from delaying other processes. • A transaction is an executing program that forms a logical unit of database processing. • includes one or more database access operations: • insertion, deletion, modification, or retrieval operations. • read_item(X). 1. Find the address of the disk block that contains item X. 2. Copy that disk block into a buffer in main memory (if that disk block is not already in some main memory buffer). 3. Copy item X from the buffer to the program variable named X. • write_item(X). 1. Find the address of the disk block that contains item X. 2. Copy that disk block into a buffer in main memory (if that disk block is not already in some main memory buffer). 3. Copy item X from the program variable named X into its correct location in the buffer. 4. Store the updated block from the buffer back to disk (either immediately or at some later point in time). • Interactions among concurrently executing transactions can cause the database state to become inconsistent. • Timing of individual steps of different transactions needs to be regulated in some manner. • This regulation is the job of the scheduler component of the DBMS, and the general process of assuring that transactions preserve consistency when executing simultaneously is called concurrency control.
• The scheduler takes read/write
requests from transactions and either executes them in buffers or delays them • BEGIN_TRANSACTION • READ or WRITE • END_TRANSACTION • COMMIT_TRANSACTION • ROLLBACK (or ABORT) • Oracle REDO and UNDO • https://www.youtube.com/watch?v=X9mTrflGhvU • Oracle Architecture • https://www.youtube.com/watch?v=bt8-QJr8bqA • The System Log • Log or Journal: The log keeps track of all transaction operations that affect the values of database items. • This information may be needed to permit recovery from transaction failures. • The log is kept on disk, so it is not affected by any type of failure except for disk or catastrophic failure. • In addition, the log is periodically backed up to archival storage (tape) to guard against such catastrophic failures. • The System Log (cont): • T in the following discussion refers to a unique transaction-id that is generated automatically by the system and is used to identify each transaction: • Types of log record: • [start_transaction,T]: Records that transaction T has started execution. • [write_item,T,X,old_value,new_value]: Records that transaction T has changed the value of database item X from old_value to new_value. • [read_item,T,X]: Records that transaction T has read the value of database item X. • [commit,T]: Records that transaction T has completed successfully, and affirms that its effect can be committed (recorded permanently) to the database. • [abort,T]: Records that transaction T has been aborted. Recovery using log records: • If the system crashes, we can recover to a consistent database state by examining the log and using one of the techniques described in later slides. 1. Because the log contains a record of every write operation that changes the value of some database item, it is possible to undo the effect of these write operations of a transaction T by tracing backward through the log and resetting all items changed by a write operation of T to their old_values. 2. We can also redo the effect of the write operations of a transaction T by tracing forward through the log and setting all items changed by a write operation of T (that did not get done permanently) to their new_values. Commit Point of a Transaction: • Definition a Commit Point: • A transaction T reaches its commit point when all its operations that access the database have been executed successfully and the effect of all the transaction operations on the database has been recorded in the log. • Beyond the commit point, the transaction is said to be committed, and its effect is assumed to be permanently recorded in the database. • The transaction then writes an entry [commit,T] into the log. • Roll Back of transactions: • Needed for transactions that have a [start_transaction,T] entry into the log but no commit entry [commit,T] into the log. Commit Point of a Transaction (cont): • Redoing transactions: • Transactions that have written their commit entry in the log must also have recorded all their write operations in the log; otherwise, they would not be committed, so their effect on the database can be redone from the log entries. Notice that the log file must be kept on disk. • At the time of a system crash, only the log entries that have been written back to disk are considered in the recovery process because the contents of main memory may be lost. • Force writing a log: • Before a transaction reaches its commit point, any portion of the log that has not been written to the disk yet must now be written to the disk. • This process is called force-writing the log file before committing a transaction • “Correctness Principle” : every transaction, if executed in isolation (without any other transactions running concurrently), will transform any consistent state to another consistent state. • In practice, transactions often run concurrently with other transactions, so the correctness principle doesn’t apply directly. • This section introduces the notion of “schedules,” the sequence of actions performed by transactions and “serializable schedules,” which produce the same result as if the transactions executed one-at-a-time. • A schedule is a sequence of the important actions taken by one or more transactions. • When studying concurrency control, the important read and write actions take place in the main-memory buffers, not the disk. • That is, a database element A that is brought to a buffer by some transaction T may be read or written in that buffer not only by T but by other transactions that access A. • Let us consider two transactions and the effect on the database when their actions are executed in certain orders. The important actions of the transactions T1 and T2 are shown in Fig. The variables t and s are local variables of T1 and T2, respectively; they are not database elements. We shall assume that the only consistency constraint on the database state is that A = B. Since T1 adds 100 to both A and B , and T2 multiplies both A and B by 2, we know that each transaction, run in isolation, will preserve consistency. • A schedule is serial if its actions consist of all the actions of one transaction, then all the actions of another transaction, and so on. No mixing of the actions is allowed. • In general, we would not expect the final state of a database to be independent of the order of transactions.
Serial schedule in which T1 precedes T2 Serial schedule in which T2 precedes T1
• Are there any other schedules that also are guaranteed to preserve consistency? • In general, we say a schedule S is serializable if there is a serial schedule S' such that for every initial database state, the effects of S and S‘ are the same.
A serializable, but not serial, schedule A nonserializable schedule
• To make the notation precise: • An action is an expression of the form ri(X ) or wi(X), meaning that transaction Ti reads or writes, respectively, the database element X . • A transaction Ti is a sequence of actions with subscript i. • A schedule S of a set of transactions T is a sequence of actions, in which for each transaction Ti in ƒ, the actions of Ti appear in S in the same order that they appear in the definition of Tj itself. We say that S is an interleaving of the actions of the transactions of which it is composed. • Schedulers in commercial systems generally enforce a condition, called “conflict serializability,” that is stronger than the general notion of serializability. • It is based on the idea of a conflict: a pair of consecutive actions in a schedule such that, if their order is interchanged, then the behavior of at least one of the transactions involved can change. • To begin, let us observe that most pairs of actions do not conflict. In what follows, we assume that Ti , and TJ are different transactions; i.e., i != J. • ri (X); rJ(Y) is never a conflict, • ri(X); wJ(Y) is not a conflict provided X != Y. • wi(X); rJ(Y) is not a conflict if X != Y, • wi(X); wJ(Y) is not a conflict as long as X != Y. • There are three situations where we may not swap the order of actions: • ri(X); wi(Y), always conflict. • wi(X); wJ(X) is a conflict. • ri(X); wJ(X) is a conflict, and so is wi(X); rJ(X). • The conclusion we draw is that any two actions of different transactions may be swapped unless: • They involve the same database element, and • At least one is a write. • We may take any schedule and make as many nonconflicting swaps as we wish, with the goal of turning the schedule into a serial schedule. • If we can do so, then the original schedule is serializable, because its effect on the database state remains the same as we perform each of the nonconflicting swaps. • We claim this schedule is conflict-serializable. Figure shows the sequence of swaps in which this schedule is converted to the serial schedule (T1, T2), where all of T1’s actions precede all those of T2. We have underlined the pair of adjacent actions about to be swapped at each step.
Converting a conflict-serializable schedule to a serial schedule by
swaps of adjacent actions • Testing Conflict Serializability of a Schedule S • For each transaction Ti participating in schedule S, create a node labeled Ti in the precedence graph. • For each case in S where TJ executes a read_item(X) after Ti executes a write_item(X), create an edge (Ti → TJ) in the precedence graph. • w i (X), r J (X) → (Ti → TJ) • For each case in S where TJ executes a write_item(X) after Ti executes a read_item(X), create an edge (Ti → TJ) in the precedence graph. • r i (X), w J (X) → (Ti → TJ) • For each case in S where TJ executes a write_item(X) after Ti executes a write_item(X), create an edge (Ti → TJ) in the precedence graph. • w i (X), w J (X) → (Ti → TJ) • The schedule S is serializable if and only if the precedence graph has no cycles. Examples of serial and nonserial schedules involving transactions T1 and T2. (a) Serial schedule A: T1 followed by T2. (b) Serial schedule B: T2 followed by T1. (c) Two nonserial schedules C and D with interleaving of operations.
Constructing the precedence graphs for schedules
A to D from Figure to test for conflict serializability. (a) Precedence graph for serial schedule A. (b) Precedence graph for serial schedule B. (c) Precedence graph for schedule C (not serializable). (d) Precedence graph for schedule D (serializable, equivalent to schedule A). • For the following schedule: • What is the precedence graph for the schedule? • Is the schedule conflict-serializable? If so, what are all the equivalent serial schedules?
• The precedence graph is:
• T3 -----> T2 -----> T1 • The schedule is thus conflict-serializable. The only conflict-equivalent serial schedule is (T3, T2, T1). • Transactions, Schedules, Serial and Serializable Schedules • We may take any schedule and make as many nonconflicting swaps as we wish, with the goal of turning the schedule into a serial schedule. • They involve the same database element, and • At least one is a write. • Precedence Graph • w i (X), r J (X) → (Ti → TJ) • r i (X), w J (X) → (Ti → TJ) • w i (X), w J (X) → (Ti → TJ) • S = r1 (X); r2 (Z); r1 (Z); r3 (X); r3 (Y); w1 (X); c1; w3 (Y); c3; r2 (Y); w2 (Z); w2 (Y); c2; • Recoverable? • TJ commits after Ti if TJ has read any data item written by Ti. • Note : Ci > C J means Ci happens before C J. Ai denotes abort Ti. To test if a schedule is recoverable one has to include abort operations. Thus in testing the recoverability, abort operations will have to be used in place of commit one at a time. Also the strictest condition is where a transaction neither reads nor writes to a data item, which was written to by a transaction that has not committed yet. • If A1>C3>C2, then S is recoverable because rolling back of T1 does not affect T2 and T3. • If C1>A3>C2, S is not recoverable because T2 read the value of Y (r2 (Y)) after T3 wrote Y (w3 (Y)) and T2 committed but T3 rolled back. Thus, T2 used non- existent value of Y. • If C1>C 3 >A 2, then S is recoverable because roll back of T2 does not affect T1 and T3. Strictest condition of S is C3>C2. • “Locks” are maintained on database elements to prevent unserializable behavior. • Intuitively, a transaction obtains locks on the database elements it accesses to prevent other transactions from accessing these elements at roughly the same time and thereby incurring the risk of unserializability. • In Fig. Below we see a scheduler that uses a lock table to help perform its job. • Responsibility of the scheduler is to take requests from transactions and either allow them to operate on the database or block the transaction until such time as it is safe to allow it to continue. • A lock table will be used to guide this decision. • li(X): Transaction Ti requests a lock on database element X . • ui(X): Transaction Ti releases (“unlocks”) its lock on database element X. • The consistency condition for transactions can be stated as: “Whenever a transaction Ti has an action ri(X) or wi(X), then there is a previous action li(X) with no intervening action ui(X), and there is a subsequent ui(X).” • The legality of schedules is stated: “If there are actions li(X) followed by lJ(X) in a schedule, then somewhere between these actions there must be an action ui(X). • The schedule is legal because the two transactions never hold a lock on A at the same time, and likewise for B. • Specifically, T2 does not execute l2(A) until after T1 executes u1(A), and T1 does not execute l1(B) until after T2 executes u2(B). • As we see from the trace of the values computed, the schedule, although legal, is not serializable. • Here are T1 and T2 from same example, with simple but important changes, in which T1 and T2 each lock B before releasing the lock on A. • When T2 requests a lock on B, the scheduler must deny the lock, because T1 still holds a lock on B. Thus, T2 is delayed, and the next actions are from T1.
Eventually, T1 executes u1(A), which
unlocks B. Now, T2 can get its lock on B, which is executed at the next step. Notice that because T2 was forced to wait, it wound up multiplying B by 2 after T1 added 100, resulting in a consistent database state. • A transaction is said to follow the two-phase locking protocol if all locking operations (read_lock, write_lock) precede the first unlock operation in the transaction. • Such a transaction can be divided into two phases: • an expanding or growing (first) phase, during which new locks on items can be acquired but none can be released; • and a shrinking (second) phase, during which existing locks can be released but no new locks can be acquired. • The main problem is that a transaction T must take a lock on a database element X even if it only wants to read X and not write it. • We cannot avoid taking the lock, because if we didn’t, then another transaction might write a new value for X while T was active and cause unserializable behavior. • On the other hand, there is no reason why several transactions could not read X at the same time, as long as none is allowed to write X . • The most common locking scheme, where there are two different kinds of locks, • one for reading (called a “shared lock” or “read lock”), • and one for writing (called an “exclusive lock” or “write lock”). • For any database element X there can be either one exclusive lock on X, or no exclusive locks but any number of shared locks. • If we want to write X , we need to have an exclusive lock on X, but if we wish only to read X we may have either a shared or exclusive lock on X. • If we want to read X but not write it, it is better to take only a shared lock. • We shall use sli(X) to mean “transaction Ti requests a shared lock on database element X ” and xli(X) for “Ti requests an exclusive lock on X.” • We continue to use ui(X) to mean that Ti unlocks X; i.e., it relinquishes whatever lock(s) it has on X . • The three kinds of requirements - consistency and 2PL for transactions, and legality for schedules - each have their counterpart for a shared/exclusive lock system. • Consistency of transactions: • A read action ri(X) must be preceded by sli(X) or xli(X), with no intervening ui(X). • A write action wi(X) must be preceded by xli(X), with no intervening ui(X). All locks must be followed by an unlock of the same element. • Two-phase locking of transactions: Locking must precede unlocking. To be more precise, in any two-phase locked transaction Ti, no action sli(X) or xli(X) can be preceded by an action ui(Y), for any Y. • Legality of schedules: An element may either be locked exclusively by one transaction or by several in shared mode, but not both. More precisely: • If xli(X) appears in a schedule, then there cannot be a following xlJ(X) or slJ(X), for some J other than i, without an intervening ui(X). • If sli(X) appears in a schedule, then there cannot be a following xlJ(X), for J != i, without an intervening ui(X). • Both T1 and T2 read A and B, but only T1 writes B. Neither writes A. • In Fig. below is an interleaving of the actions of T1 and T2 in which T1 begins by getting a shared lock on A. • Then, T2 follows by getting shared locks on both A and B. Now, T1 needs an exclusive lock on B, since it will both read and write B.
• However, it cannot get the
exclusive lock because T2 already has a shared lock on B. • Thus, the scheduler forces T1 to wait. • Eventually, T2 releases the lock on B. At that time, T1 may complete. • If we use several lock modes, then the scheduler needs a policy about when it can grant a lock request, given the other locks that may already be held on the same database element. • A compatibility matrix is a convenient way to describe lock-management policies. • Deadlock occurs when each transaction T in a set of two or more transactions is waiting for some item that is locked by some other transaction T’ in the set. • Hence, each transaction in the set is in a waiting queue, waiting for one of the other transactions in the set to release the lock on an item. • But because the other transaction is also waiting, it will never release the lock. • The simplest way to detect and resolve deadlocks is with a timeout. Put a limit on how long a transaction may be active, and if a transaction exceeds this time, roll it back. • For example, in a simple transaction system, where typical transactions execute in milliseconds, a timeout of one minute would affect only transactions that are caught in a deadlock. • Deadlocks that are caused by transactions waiting for locks held by another can be detected by a waits-for graph, indicating which transactions are waiting for locks held by another transaction. • This graph can be used either to detect deadlocks after they have formed or to prevent deadlocks from ever forming. • The waits-for graph has a node for each transaction that currently holds any lock or is waiting for one. There is an arc from node (transaction) T to node U if there is some database element • A such that: • U holds a lock on A, • T is waiting for a lock on A, and • T cannot get a lock on A in its desired mode unless U first releases its lock on A • If there are no cycles in the waits-for graph, then each transaction can complete eventually. •
Waits-for Waits-for graph with a
graph after cycle caused by step step (7) (8) • An alternative to maintaining the waits-for graph is to associate with each transaction a timestamp. • The timestamp is used when a transaction T has to wait for a lock that is held by another transaction U. • Two different things happen, depending on whether T or U is older (has the earlier timestamp). • There are two different policies that can be used to manage transactions and detect deadlocks. • The Wait-Die Scheme: • If T is older than U (i.e., the timestamp of T is smaller than U’s timestamp), then T is allowed to wait for the lock(s) held by U. • If U is older than T, then T “dies”; it is rolled back. • The Wound-Wait Scheme: • If T is older than U, it “wounds” U. Usually, the “wound” is fatal: U must roll back and relinquish to T the lock(s) that T needs from U. There is an exception if, by the time the “wound” takes effect, U has already finished and released its locks. In that case, U survives and need not be rolled back. • If U is older than T, then T waits for the lock(s) held by U. • THE END