作者: Parrotao

orace的隔離級(jí)別

隔離級(jí)別（isoation eve）

隔離級(jí)別定義了事務(wù)與事務(wù)之間的隔離程度。

隔離級(jí)別與并發(fā)性是互為矛盾的：隔離程度越高，數(shù)據(jù)庫(kù)的并發(fā)性越差；隔離程度越低，數(shù)據(jù)庫(kù)的并發(fā)性越好。

ANSI/ISO SQ92標(biāo)準(zhǔn)定義了一些數(shù)據(jù)庫(kù)操作的隔離級(jí)別：

• 未提交讀（read uncommitted）
• 提交讀（read committed）  
• 重復(fù)讀（repeatabe read）  
• 序列化（seriaizabe）

通過(guò)一些現(xiàn)象，可以反映出隔離級(jí)別的效果。這些現(xiàn)象有：

• 更新丟失（ost update）：當(dāng)系統(tǒng)允許兩個(gè)事務(wù)同時(shí)更新同一數(shù)據(jù)是，發(fā)生更新丟失。  
• 臟讀（dirty read）：當(dāng)一個(gè)事務(wù)讀取另一個(gè)事務(wù)尚未提交的修改時(shí)，產(chǎn)生臟讀。
• 非重復(fù)讀（nonrepeatabe read）：同一查詢?cè)谕皇聞?wù)中多次進(jìn)行，由于其他提交事務(wù)所做的修改或刪除，每次返回不同的結(jié)果集，此時(shí)發(fā)生非重復(fù)讀。(A transaction rereads data it has previousy read and finds that another committed transaction has modified or deeted the data.  )
• 幻像（phantom read）：同一查詢?cè)谕皇聞?wù)中多次進(jìn)行，由于其他提交事務(wù)所做的插入操作，每次返回不同的結(jié)果集，此時(shí)發(fā)生幻像讀。(A transaction reexecutes a query returning a set of rows that satisfies a search condition and finds that another committed transaction has inserted additiona rows that satisfy the condition.  )

下面是隔離級(jí)別及其對(duì)應(yīng)的可能出現(xiàn)或不可能出現(xiàn)的現(xiàn)象

ORACE的隔離級(jí)別

ORACE提供了SQ92標(biāo)準(zhǔn)中的read committed和seriaizabe，同時(shí)提供了非SQ92標(biāo)準(zhǔn)的read-ony。

read committed：

• 這是ORACE缺省的事務(wù)隔離級(jí)別。
• 事務(wù)中的每一條語(yǔ)句都遵從語(yǔ)句級(jí)的讀一致性。
• 保證不會(huì)臟讀；但可能出現(xiàn)非重復(fù)讀和幻像。

seriaizabe：（串行執(zhí)行事務(wù)，并發(fā)性最小）

• 簡(jiǎn)單地說(shuō)，seriaizabe就是使事務(wù)看起來(lái)象是一個(gè)接著一個(gè)地順序地執(zhí)行。
• 僅僅能看見(jiàn)在本事務(wù)開(kāi)始前由其它事務(wù)提交的更改和在本事務(wù)中所做的更改。
• 保證不會(huì)出現(xiàn)非重復(fù)讀和幻像。
• Seriaizabe隔離級(jí)別提供了read-ony事務(wù)所提供的讀一致性（事務(wù)級(jí)的讀一致性），同時(shí)又允許DM操作。

如果有在seriaizabe事務(wù)開(kāi)始時(shí)未提交的事務(wù)在seriaizabe事務(wù)結(jié)束之前修改了seriaizabe事務(wù)將要修改的行并進(jìn)行了提交，則seriaizabe事務(wù)不會(huì)讀到這些變更，因此發(fā)生無(wú)法序列化訪問(wèn)的錯(cuò)誤。（換一種解釋方法：只要在seriaizabe事務(wù)開(kāi)始到結(jié)束之間有其他事務(wù)對(duì)seriaizabe事務(wù)要修改的東西進(jìn)行了修改并提交了修改，則發(fā)生無(wú)法序列化訪問(wèn)的錯(cuò)誤。）

If a serializable transaction contains data manipulation language (DML) that attempts to update any resource that may have been updated in a transaction uncommitted at the start of the serializable transaction, （并且修改在后來(lái)被提交而沒(méi)有回滾），then the DML statement fails. 返回的錯(cuò)誤是ORA-08177: Cannot serialize access for this transaction。

ORACE在數(shù)據(jù)塊中記錄最近對(duì)數(shù)據(jù)行執(zhí)行修改操作的N個(gè)事務(wù)的信息，目的是確定本事務(wù)開(kāi)始時(shí)，是否存在未提交的事務(wù)修改了本事務(wù)將要修改的行。具體見(jiàn)英文：

Oracle permits a serializable transaction to modify a data row only if it can determine that prior changes to the row were made by transactions that had committed when the serializable transaction began.

To make this determination efficiently, Oracle uses control information stored in the data block that indicates which rows in the block contain committed and uncommitted changes. In a sense, the block contains a recent history of transactions that affected each row in the block. The amount of history that is retained is controlled by the INITRANS parameter of CREATE TABLE and ALTER TABLE. Under some circumstances, Oracle may have insufficient history information to determine whether a row has been updated by a "too recent" transaction. This can occur when many transactions concurrently modify the same data block, or do so in a very short period. You can avoid this situation by setting higher values of INITRANS for tables that will experience many transactions updating the same blocks. Doing so will enable Oracle to allocate sufficient storage in each block to record the history of recent transactions that accessed the block.

The INITRANS Parameter：Oracle stores control information in each data block to manage access by concurrent transactions. Therefore, if you set the transaction isolation level to serializable, you must use the ALTER TABLE command to set INITRANS to at least 3. This parameter will cause Oracle to allocate sufficient storage in each block to record the history of recent transactions that accessed the block. Higher values should be used for tables that will undergo many transactions updating the same blocks.

read-ony：

• 遵從事務(wù)級(jí)的讀一致性，僅僅能看見(jiàn)在本事務(wù)開(kāi)始前由其它事務(wù)提交的更改。
• 不允許在本事務(wù)中進(jìn)行DM操作。
• read ony是seriaizabe的子集。它們都避免了非重復(fù)讀和幻像。區(qū)別是在read ony中是只讀；而在seriaizabe中可以進(jìn)行DM操作。
• Export with CONSISTENT = Y sets the transaction to read-ony.

read committed和seriaizabe的區(qū)別和聯(lián)系：

事務(wù)1先于事務(wù)2開(kāi)始，并保持未提交狀態(tài)。事務(wù)2想要修改正被事務(wù)1修改的行。事務(wù)2等待。如果事務(wù)1回滾，則事務(wù)2（不論是read committed還是seriaizabe方式）進(jìn)行它想要做的修改。如果事務(wù)1提交，則當(dāng)事務(wù)2是read committed方式時(shí)，進(jìn)行它想要做的修改；當(dāng)事務(wù)2是seriaizabe方式時(shí)，失敗并報(bào)錯(cuò)“Cannot seriaize access”，因?yàn)槭聞?wù)2看不見(jiàn)事務(wù)1提交的修改，且事務(wù)2想在事務(wù)1修改的基礎(chǔ)上再做修改。

即seriaizabe不允許存在事務(wù)嵌套

具體見(jiàn)英文：

Both read committed and serializable transactions use row-level locking, and both will wait if they try to change a row updated by an uncommitted concurrent transaction. The second transaction that tries to update a given row waits for the other transaction to commit or roll back and release its lock. If that other transaction rolls back, the waiting transaction (regardless of its isolation mode) can proceed to change the previously locked row, as if the other transaction had not existed. However, if the other (blocking) transaction commits and releases its locks, a read committed transaction proceeds with its intended update. A serializable transaction, however, fails with the error "Cannot serialize access", because the other transaction has committed a change that was made since the serializable transaction began.

read committed和seriaizabe可以在ORACE并行服務(wù)器中使用。

關(guān)于SET TRANSACTION READ WRITE：read write和read committed 應(yīng)該是一樣的。在讀方面，它們都避免了臟讀，但都無(wú)法實(shí)現(xiàn)重復(fù)讀。雖然沒(méi)有文檔說(shuō)明read write在寫(xiě)方面與read committed一致，但顯然它在寫(xiě)的時(shí)候會(huì)加排他鎖以避免更新丟失。在加鎖的過(guò)程中，如果遇到待鎖定資源無(wú)法鎖定，應(yīng)該是等待而不是放棄。這與read committed一致。

語(yǔ)句級(jí)的讀一致性

• ORACE保證語(yǔ)句級(jí)的讀一致性，即一個(gè)語(yǔ)句所處理的數(shù)據(jù)集是在單一時(shí)間點(diǎn)上的數(shù)據(jù)集，這個(gè)時(shí)間點(diǎn)是這個(gè)語(yǔ)句開(kāi)始的時(shí)間。
• 一個(gè)語(yǔ)句看不見(jiàn)在它開(kāi)始執(zhí)行后提交的修改。
• 對(duì)于DM語(yǔ)句，它看不見(jiàn)由自己所做的修改，即DM語(yǔ)句看見(jiàn)的是它本身開(kāi)始執(zhí)行以前存在的數(shù)據(jù)。

事務(wù)級(jí)的讀一致性

• 事務(wù)級(jí)的讀一致性保證了可重復(fù)讀，并保證不會(huì)出現(xiàn)幻像。

設(shè)置隔離級(jí)別

設(shè)置一個(gè)事務(wù)的隔離級(jí)別

• SET TRANSACTION ISOLATION LEVEL READ COMMITTED;
• SET TRANSACTION ISOLATION LEVEL SERIALIZABLE;
• SET TRANSACTION READ ONLY;

設(shè)置增個(gè)會(huì)話的隔離級(jí)別

• ATER SESSION SET ISOLATION_LEVE SERIALIZABLE;
• ATER SESSION SET ISOLATION_LEVE READ COMMITTED;

Choice of Isolation Level

Application designers and developers should choose an isolation level based on application performance and consistency needs as well as application coding requirements.

For environments with many concurrent users rapidly submitting transactions, designers must assess transaction performance requirements in terms of the expected transaction arrival rate and response time demands. Frequently, for high-performance environments, the choice of isolation levels involves a trade-off between consistency and concurrency.

Read Committed Isolation

For many applications, read committed is the most appropriate isolation level. Read committed isolation can provide considerably more concurrency with a somewhat increased risk of inconsistent results due to phantoms and non-repeatable reads for some transactions.

Many high-performance environments with high transaction arrival rates require more throughput and faster response times than can be achieved with serializable isolation. Other environments that supports users with a very low transaction arrival rate also face very low risk of incorrect results due to phantoms and nonrepeatable reads. Read committed isolation is suitable for both of these environments.

兩種情況：（1）在事務(wù)量大、高性能的計(jì)算環(huán)境，需要更高的吞吐量和響應(yīng)時(shí)間；（2）事務(wù)數(shù)少，并且發(fā)生幻影和不可重復(fù)讀的幾率的比較低

Oracle read committed isolation provides transaction set consistency for every query. That is, every query sees data in a consistent state. Therefore, read committed isolation will suffice for many applications that might require a higher degree of isolation if run on other database management systems that do not use multiversion concurrency control.

Read committed isolation mode does not require application logic to trap the "Cannot serialize access" error and loop back to restart a transaction. In most applications, few transactions have a functional need to issue the same query twice, so for many applications protection against phantoms and non-repeatable reads is not important. Therefore many developers choose read committed to avoid the need to write such error checking and retry code in each transaction.

Serializable Isolation

Oracle's serializable isolation is suitable for environments where there is a relatively low chance that two concurrent transactions will modify the same rows and the long-running transactions are primarily read-only. It is most suitable for environments with large databases and short transactions that update only a few rows.

（1）適合于很少存在兩個(gè)事務(wù)同時(shí)修改同一條記錄的情況

（2）長(zhǎng)事務(wù)以只讀為主

（3）大型數(shù)據(jù)庫(kù)并且每個(gè)短事務(wù)只修改很少的記錄

Serializable isolation mode provides somewhat more consistency by protecting against phantoms and nonrepeatable reads and can be important where a read/write transaction executes a query more than once.

Unlike other implementations of serializable isolation, which lock blocks for read as well as write, Oracle provides nonblocking queries and the fine granularity of row-level locking, both of which reduce write/write contention. For applications that experience mostly read/write contention, Oracle serializable isolation can provide significantly more throughput than other systems. Therefore, some applications might be suitable for serializable isolation on Oracle but not on other systems.

All queries in an Oracle serializable transaction see the database as of a single point in time, so this isolation level is suitable where multiple consistent queries must be issued in a read/write transaction. A report-writing application that generates summary data and stores it in the database might use serializable mode because it provides the consistency that a READ ONLY transaction provides, but also allows INSERT, UPDATE, and DELETE.

作者聯(lián)系方式:

E     TaoPuyin
G   System Engineer Oracle DBA
G  Fuji Xerox China Limited (Shanghai)