关于Oracle Rac数据不同步
欢迎进入Oracle社区论坛,与200万技术人员互动交流 >>进入 现在有这样的环境: 一台web Server,一个是纯JAVA APP 程序 数据库两台做成RAC的形式。 web Server与APP 程序都通过oci(rac)的方式连接 数据库。 出了这样的怪问题,webServer更新或是插图入一条数
欢迎进入Oracle社区论坛,与200万技术人员互动交流 >>进入
现在有这样的环境:
一台web Server,一个是纯JAVA APP 程序
数据库两台做成RAC的形式。 web Server与APP 程序都通过oci(rac)的方式连接
数据库。
出了这样的怪问题,webServer更新或是插图入一条数据,后面紧跟着的在APP中就查询不到,等到用工具查询就没有问题.
初步怀疑
1. RAC方式下面的数据库两个instance的同步没做好?
查询相关资料发现在与MAX_COMMIT_PROPAGATION_DELAY有关.
最大提交传播时延(MAX_COMMIT_PROPAGATION_DELAY,简称MCPD),在ORACLE RAC(或OPS)环境中才使用,表示在RAC系统中,一个instance系统提交产生的最新系统改变码(SCN),能够以多快的速度反应到另一个instance中。
举例说明,RAC系统,有A,B两个实例(instance),A、B本地系统改变码为SCN1,A更新数据DATA1提交, LGWR操作完成后,A本地系统改变码为SCN2,经过不大于MAX_COMMIT_PROPAGATION_DELAY时间后,B系统本地改变码才变为SCN2。
Global Cache Services 将刷新RAC中的SCN。不管SCN是否及时刷新,后续的数据查询都不会因此产生数据库错误。但,在此时间内,有可能查询结果不是最新数据,产生读一致性(read consistency)问题。
RAC环境中的所有实例,此参数值必须相同。
ORACLE8i后,建议常用的两个值是0和700(默认),其他数值皆不建议。其实,这两个数值就代表了RAC环境中,两种SCN 产生机制:
Lamport Scheme和 Broadcast on Commit scheme。
设置为默认值700,表示采用Lamport Scheme,SCN改变不会完全同步,同步将在 7秒钟内完成,而不是总等待7秒钟后才完成。如果系统比较空闲,同步可能在0.5秒(甚至更短时间)内完成;不管系统多繁忙,同步时间也不可能超过7秒。不难理解,采用此模式,整个RAC系统的运行效率较高。
设置为0,表示采用Broadcast on Commit scheme,SCN改变完全同步。每当commit时(即LGWR 写redo log时):
- LGWR发送消息更新全局SCN(global SCN),
- LGWR 发送消息给每个活动的实例更新其本地SCN(local SCN)。
有资料说,只要MCPD Lamport Scheme能够适应绝大部分应用的要求,只有个别实时性特别高的业务,才需要Broadcast on Commit scheme。通过分析,不难理解,Broadcast on Commit scheme将需要更多的系统资源

Hot AI Tools

Undresser.AI Undress
AI-powered app for creating realistic nude photos

AI Clothes Remover
Online AI tool for removing clothes from photos.

Undress AI Tool
Undress images for free

Clothoff.io
AI clothes remover

AI Hentai Generator
Generate AI Hentai for free.

Hot Article

Hot Tools

Notepad++7.3.1
Easy-to-use and free code editor

SublimeText3 Chinese version
Chinese version, very easy to use

Zend Studio 13.0.1
Powerful PHP integrated development environment

Dreamweaver CS6
Visual web development tools

SublimeText3 Mac version
God-level code editing software (SublimeText3)

Hot Topics



To query the Oracle tablespace size, follow the following steps: Determine the tablespace name by running the query: SELECT tablespace_name FROM dba_tablespaces; Query the tablespace size by running the query: SELECT sum(bytes) AS total_size, sum(bytes_free) AS available_space, sum(bytes) - sum(bytes_free) AS used_space FROM dba_data_files WHERE tablespace_

Data import method: 1. Use the SQLLoader utility: prepare data files, create control files, and run SQLLoader; 2. Use the IMP/EXP tool: export data, import data. Tip: 1. Recommended SQL*Loader for big data sets; 2. The target table should exist and the column definition matches; 3. After importing, data integrity needs to be verified.

Uninstall method for Oracle installation failure: Close Oracle service, delete Oracle program files and registry keys, uninstall Oracle environment variables, and restart the computer. If the uninstall fails, you can uninstall manually using the Oracle Universal Uninstall Tool.

Creating an Oracle table involves the following steps: Use the CREATE TABLE syntax to specify table names, column names, data types, constraints, and default values. The table name should be concise and descriptive, and should not exceed 30 characters. The column name should be descriptive, and the data type specifies the data type stored in the column. The NOT NULL constraint ensures that null values are not allowed in the column, and the DEFAULT clause specifies the default values for the column. PRIMARY KEY Constraints to identify the unique record of the table. FOREIGN KEY constraint specifies that the column in the table refers to the primary key in another table. See the creation of the sample table students, which contains primary keys, unique constraints, and default values.

There are three ways to view instance names in Oracle: use the "sqlplus" and "select instance_name from v$instance;" commands on the command line. Use the "show instance_name;" command in SQL*Plus. Check environment variables (ORACLE_SID on Linux) through the operating system's Task Manager, Oracle Enterprise Manager, or through the operating system.

Oracle View Encryption allows you to encrypt data in the view, thereby enhancing the security of sensitive information. The steps include: 1) creating the master encryption key (MEk); 2) creating an encrypted view, specifying the view and MEk to be encrypted; 3) authorizing users to access the encrypted view. How encrypted views work: When a user querys for an encrypted view, Oracle uses MEk to decrypt data, ensuring that only authorized users can access readable data.

Oracle provides multiple deduplication query methods: The DISTINCT keyword returns a unique value for each column. The GROUP BY clause groups the results and returns a non-repetitive value for each group. The UNIQUE keyword is used to create an index containing only unique rows, and querying the index will automatically deduplicate. The ROW_NUMBER() function assigns unique numbers and filters out results that contain only line 1. The MIN() or MAX() function returns non-repetitive values of a numeric column. The INTERSECT operator returns the common values of the two result sets (no duplicates).

To create a user in Oracle, follow these steps: Create a new user using the CREATE USER statement. Grant the necessary permissions using the GRANT statement. Optional: Use the RESOURCE statement to set the quota. Configure other options such as default roles and temporary tablespaces.
