ITPub博客

oracle备库延时的监控方法[@more@]根据最后一个已经apply的日志文件的next_time时间戳进行判断 select /*+ rule */ nvl(round((sysdate-max(NEXT_TIME))*24),999) from v$archived_log where APPLIED！='NO' and REGISTRAR='RFS' http://docs.oracle.com/cd/E11882_01/server.112/e25513/dynviews_1016.htm#REFRN30011 其他检查方式 1.根据system表空间文件的chkpoint时间检查 select /*+ rule */ nvl(round((sysdate-CHECKPOINT_TIME)*24),999) from v$datafile where file#=1; 2.根据最后一个apply的归档文件的创建时间判断，优点是不需

预警的些想法[@more@]1.预警的意义 通知我们某个点出现了异常，并且这个异常在可预见的未来可能会造成什么样的影响2.预警的要素 监控点&触发点&对象&渠道&频率3.一次通知定义 语音外呼-1次，短信-2条，邮件1条，旺旺1条4.常规预警分级 根据影响可能发生的时间和可能造成的影响和评估，而不是简单根据触发点的重要程度 分级的目的1） 重要性的传递2）抑制告警 分级的出发点，主要的级别为LV0，LV1，LV4，LV5。LV2,LV3的预警希望只是用做辅助性分级。 负载类：LV0，LV1，LV4，LV5 负载类告警一般突发性比较强，以短期型

Bug 9290526 - Poor plan for recursive SQL used for DML involving a UNIQUE constraint [ID 9290526.8] … [@more@] A suboptimal plan may be seen for the specific recursive SQL: select c.name, u.name from con$ c, cdef$ cd, user$ u where c.con# = cd.con# and cd.enabled =:1 and c.owner# = u.user#; This can cause inserts involving a unique constraint to show poor performance. Workaround Set optimizer_mode=rule for the affected DMLs. https://supporthtml.oracle.com/epmos/faces/ui/km/SearchDocDisplay.js

softer soft parse 软软解析[@more@]不解释，帖链接http://docs.oracle.com/cd/E11882_01/server.112/e25513/initparams228.htmhttp://docs.oracle.com/cd/E11882_01/server.112/e25513/stats002.htmhttp://www.orafaq.com/node/758http://www.orafan.net/blog/180.html

numa oracle oom[@more@]3.NUMA+oracle的一些问题 1）.rh54-hugepage oracle的sga并非完全预分配，hugepage应该是预分配出page的，虽然没有划给SGA，所以内存分配上应该不会导致问题。另外，hugepage实际分配的大多数位于node1，pagecache位于node0，虽然整体较为平均，不确定是否会有影响，不过个人偏向于影响不大。 2）.numa 进程倾向总是从同一node分配内存，对于oracle，由于使用filecache的进程以arch进程为主，且大部分时候oracle只有一个arch进程工作，可能导致filecache过分集中到一个node里面。 3）.numa

oracle11.2.0.2 dg vcs ha 切换异常[@more@]现象：oracle11.2.0.2+vcs，当有主库的public进程连接到备库时，ha无法正常完成备库实例的切换。问题关键点如下： 1）oracle11.2.0.2 shutdown immediate操作中，会等待active call状态的进程，（不确定是会等待所有处于active call状态的进程还是单单active call状态的RFS进程）；并且这个等待可能是没有超时或者超时设置相当长（从测试中看超过1h）。shutdown操作会一直卡在ipcs看到的第一个信号量（释放？）处，对应系统调用一直在报超时错误。 2）当监听器所侦听的网络异常（这里是IP被

Bug 8575528 Missing entries in V$MUTEX_SLEEP.location[@more@]Affects:Product (Component)Oracle Server (Rdbms) Range of versions believed to be affectedVersions >= 10.2.0.4 but BELOW 10.2.0.5 Versions confirmed as being affected(None Specified) Platforms affectedGeneric (all / most platforms affected) DescriptionThis problem is caused by the fix for bug 6795880. Null mutex locations may be seen in V$MUTEX_SLEEP or a core dump encountered in a call from kqlfMutexClnFetch() to strcpy().

service_name or sid[@more@]问题起源于itpub一个帖子http://www.itpub.net/thread-1358298-1-1.html，答案呢，通过service_name实现failover和lb自然是一方面，顺便查了下oracle文档相关介绍，还是有些特别的发现的。http://download.oracle.com/docs/cd/B19306_01/network.102/b14213/tnsnames.htm#sthref754.6.4.8 SERVICE_NAMEPurposeUse the SERVICE_NAME parameter to identify the Oracle9i or Oracle8 database service to access. Set the value to a value specified by the SERVICE_NAMES parameter in the initialization par

Oracle SQL performance with database links - dblink[@more@]http://download.oracle.com/docs/cd/B14117_01/server.101/b10752/hintsref.htm#30401https://support.oracle.com/CSP/main/article?cmd=show&type=NOT&doctype=PROBLEM&id=825677.1http://www.dba-oracle.com/t_sql_dblink_performance.htmhttp://jiri.wordpress.com/2009/05/20/queries-across-db-link-and-how-driving_site-hint-does-not-work-for-inserts/http://psoug.org/reference/hints.html

_optimizer_compute_index_stats[@more@]COMPUTE STATISTICS In earlier releases, you could use this clause to start or stop the collection of statistics on an index. This clause has been deprecated. Oracle Database now automatically collects statistics during index creation and rebuild. This clause is supported for backward compatibility and will not cause errors._optimizer_compute_index_stats

oracle之commit控制[@more@]_lgwr_async_io_wait_for_synccommit_writehttp://download.oracle.com/docs/cd/E11882_01/server.112/e10820/initparams032.htm#REFRN10260http://www.oracle-base.com/articles/10g/Commit_10gR2.phpcommit_waithttp://download.oracle.com/docs/cd/E11882_01/server.112/e10820/initparams031.htm#REFRN10266commit_logginghttp://download.oracle.com/docs/cd/E11882_01/server.112/e10820/initparams029.htm#REFRN10265

作者: 弦乐之花 | 可以转载, 但必须以超链接形式标明文章原始出处和作者信息及版权声明链接http://shiri512003.itpub.net/post/37713/499500[@more@]1.版本确认echo ocfs2-`uname -r` 2.下载版本http://oss.oracle.com/projects/ocfs2-tools/files/RedHat/RHEL5/x86_64/1.4.4-1/http://oss.oracle.com/projects/ocfs2/files/RedHat/RHEL5/x86_64/1.4.7-1/3.安装rpm -ivh *rpm4.命令行工具/etc/init.d/o2cbo2cb_ctlocfs2_hb_ctlcluster配置：/etc/ocfs2/cluster.confcluster:node_count = 3name = webclusternode:ip_port = 7777ip_address

作者: 弦乐之花 | 可以转载, 但必须以超链接形式标明文章原始出处和作者信息及版权声明链接http://shiri512003.itpub.net/post/37713/499390[@more@]摘自metalink：From Oracle 10G R2 , Oracle checks the options with which a NFS mount is mounted on the filesystem. and this is done to ensure that no corruption of the database can happen as incorrectly mounted NFS volumes can result in data corruption. There are no single set of NFS mount options that work across all Oracle platforms Please ensure that you ha

作者: 弦乐之花 | 可以转载, 但必须以超链接形式标明文章原始出处和作者信息及版权声明链接http://shiri512003.itpub.net/post/37713/499385[@more@]文档提到oracle 10g支持temp文件自动创建，但一直没有真正去关注这个过程，偶然的一次机会发现了这个自动创建的过程。一次open read only dg的时候，发现异常的慢，于是查看alert日志发现如下信息：Tue Apr 20 09:52:58 2010alter database open read onlyTue Apr 20 09:52:59 2010SMON: enabling cache recoveryTue Apr 20 09:53:01 2010Re-creating tempfile /opt/oracle/oradata/test/t

作者: 弦乐之花 | 可以转载, 但必须以超链接形式标明文章原始出处和作者信息及版权声明链接http://shiri512003.itpub.net/post/37713/499375[@more@]按照文档ID 399727.1说明，产生这种现象一般有两种情况：1.使用低版本的监听程序监听高版本的服务，如使用9i的监听器监听10g的服务。2.由于client端连接的不稳定性。如瞬间的大量连接则可能导致这样的问题。不过是哪种原因导致的，这时候系统都处于非健康状态，实际环境需要注意。关于该问题的详细说明请参考https://support.oracle.com/CSP/main/article?cmd=show&type=NOT&docty

How to Keep All UNDO Segments from Being Offlined in Oracle 10g - Fast Ramp-Up [ID 396863.1][@more@]先简单记录下，稍后更新How to Keep All UNDO Segments from Being Offlined in Oracle 10g - Fast Ramp-Up [ID 396863.1]Per the Oracle Database 10g New Features for Administrators Course, the concept of Fast Ramp-Up indicates the following:"In previous versions, when the instance was restarted or when you switched undo tablespaces, the ramp-up time for the system to grow the number of online undo s

http://www.orafaq.com/node/2078[@more@]转篇老外写的switch操作过来，老外的文档还是写的不错的，当然实际切换过程除了oracle本身的操作，可能还要涉及后台任务，参数文件等等的处理。以下内容转自http://www.orafaq.com/node/2078Oracle 10g Data Guard – Quick Switchover with Physical Standby DatabaseSubmitted by hjiao on Mon, 2008-09-22 09:14 RDBMS ServerOnce the standby database is setup using Data Guard and works properly, you may want to test switchover, or perform switchover to reduce primary database do

作者: 弦乐之花 | 可以转载, 但必须以超链接形式标明文章原始出处和作者信息及版权声明链接http://shiri512003.itpub.net/post/37713/497426[@more@]在自动PGA管理模式下，下面的统计信息对于我们理解PGA的整体使用情况是很有帮助的。记录下文档关于这个视图的描述。http://download.oracle.com/docs/cd/B28359_01/server.111/b28320/dynviews_2093.htm#REFRN30180Statistic NameDescriptionaggregate PGA target parameterCurrent value of the PGA_AGGREGATE_TARGET initialization parameter. If this parameter is not set, then its

作者: 弦乐之花 | 可以转载, 但必须以超链接形式标明文章原始出处和作者信息及版权声明 网址http://shiri512003.itpub.net/post/37713/496826[@more@]pub网友提出的一个问题：目前生产系统有一个ASM的DB，这个ASM对应一个DISKGROUP ASM01，对应存储上的四个LUN,现在想做一个FAILGROUP,现在另外挂上去了4个LUN，并且建立一个DISKGROUP ASM02，用什么命令可以让这两个DISKGROUP建立FAILGROUP关系呢?帖子链接http://www.itpub.net/thread-1268220-1-1.html查了文档，没看到更改diskgroup冗余度的命令，没有向diskgroup中add failgroup的命令1

作者: 弦乐之花 | 可以转载, 但必须以超链接形式标明文章原始出处和作者信息及版权声明 网址http://shiri512003.itpub.net/post/37713/496824[@more@]这个问题首先由eagle提出，相关链接为http://www.dbafan.com/blog/?p=295#comment-38424这里只是自己在10204linux平台上测试一下相关脚本-- 脚本取自tom 9i&10g编程艺术P172-173-- reset_stat.sqldrop table sess_stats;create table sess_stats( name varchar2(64), value number, diff number );variable sid numberexec :sid := &1-- watch_stat.sqlmerge into sess_statsusin