http://watershed.itpub.net Sun,21 03 2010 16:06:10 http://www.plogworld.net 最近在一个客户那里安装9iforAIX，结果运行runInstaller报错如下：security properties not found. using defaults.stackpointer=42ba1f70Writing Java core file ....Written Java core to /tmp/javacore434372.1262950584.txt/tmp/OraInstall2010-01-08_07-36-16PM/installer: 434372 Memory fault(coredump)在metalink没有找到一样环境一样报错的情况，于是检查了很多相关的设置，包括操作系统的补丁和内核等等，没有发现问题。只可能是安装介质的问题了，结果重新上传了安装介质，安装成功。 http://watershed.itpub.net/post/365/497470 http://watershed.itpub.net/post/365/497470 http://watershed.itpub.net/post/365/497470 watershed Tech Documents Sun,14 03 2010 23:08:36 zhangsy之家 1. REMOVING A NODE FROM A 10g RAC CLUSTERAssume you want to delete node 2 from cluster1> srvctl stop nodeapps -n (run on delete node)2> run netca to delete listener that on the delete node3> srvctl remove nodeapps -n (run on delete node) 4> on remaining node : run the installer with theupdateNodeList option ,you can run this command with local_node=host_name option just when the command run unsuccessfullyexport DISPLAY=XXX.XXX.XXX.XXX：X.0$ORACLE_HOME/oui/bin/runInstaller -updateNodeListORACLE_H... http://watershed.itpub.net/post/365/497182 http://watershed.itpub.net/post/365/497182 http://watershed.itpub.net/post/365/497182 watershed Tech Documents Thu,04 03 2010 23:13:40 zhangsy之家 今天晚上在贵州电信升级一套rac，发生异常，数据库升级到9208以后无法以cluster的方式启动，在metalink查了好多文档，都没有能够匹配的，见了鬼了，最终经过5个小时折腾终于找到了答案。环境：oracle9207 rac 升级到oracle9208 rac问题现象：升级完成后两个节点的数据库无法同时启动，只能启动一个，无论先启动哪个节点，另外一个节点都无法启动，在mount阶段hang报错：node1：Sat Jan 9 03:13:27 2010starting up 1 shared server(s) ...starting up 1 dispatcher(s) for network address '(ADDRESS=(PARTIAL=YES)(PROTOCOL=TCP))'...ARC... http://watershed.itpub.net/post/365/495740 http://watershed.itpub.net/post/365/495740 http://watershed.itpub.net/post/365/495740 watershed Tech Documents Sat,09 01 2010 05:10:07 zhangsy之家 ZT: http://yangtingkun.itpub.net/测试一Oracle的RAC的高可用功能除了负载均衡还包括TAF（Transparent Application Failover）。RAC的TAF是指会话连接到一个实例上，如果这个实例出现了故障，Oracle会自动将会话迁移到另一个实例上。看一个简单的例子。首先不配置TAF，在客户端TNSNAMES.ORA中进行如下的测试：TESTRAC =(DESCRIPTION =(ADDRESS = (PROTOCOL = TCP)(HOST = 172.25.198.224)(PORT = 1521))(ADDRESS = (PROTOCOL = TCP)(HOST = 172.25.198.225)(PORT = 1521))(CONNECT_DATA =(SERVER = DEDICATED)(SERVICE_NAME = testrac)))... http://watershed.itpub.net/post/365/467679 http://watershed.itpub.net/post/365/467679 http://watershed.itpub.net/post/365/467679 watershed Tech Documents Wed,30 07 2008 11:25:21 zhangsy之家 在statspack中有大量的global cache cr request:这个等待事件主要是节点间通讯异常造成的.1.需要确定应用的分布,尽量将不同的业务类型分配到不同的instance上面2. 确定inter connect网络状态[root]# netstat -p udpudp:3500183652 datagrams received0 incomplete headers0 bad data length fields0 bad checksums8402728 dropped due to no socket12017785 broadcast/multicast datagrams dropped due to no socket231620 socket buffer overflows3479531519 delivered3119249970 datagrams output如果存在较大的socket buffer overflows... http://watershed.itpub.net/post/365/467615 http://watershed.itpub.net/post/365/467615 http://watershed.itpub.net/post/365/467615 watershed Tech Documents Tue,29 07 2008 15:50:56 zhangsy之家 create a physical standby database steps:1.physical standby 是将primary database的archielog apply 到standby database,从而完成primary database 和 standby database的数据一致性2.standby 有以下先决条件必须满足:2.1 oracle的版本必须一致2.2 操作系统必须一致(版本没有要求)2.3 primary database 必须运行在archive mode下2.4 数据库文件结构和名称必须一致,如果不一致,则必须在standby machine 做链接文件2.5 primary database必须 force logging mode3.下面以primary 和standby 分部在不同host上为例:3.1 确认primary 运行在a... http://watershed.itpub.net/post/365/457140 http://watershed.itpub.net/post/365/457140 http://watershed.itpub.net/post/365/457140 watershed Tech Documents Thu,13 03 2008 11:47:46 zhangsy之家 站在窗前,我看着她慢慢的消失在我的视线里,我哭了.我一直看着她消失的地方,很久.我心里只有内疚.我不知道该说什么好,只有"对不起".在我25岁的时候我就觉得我不会再为女孩哭了,可是这次我还是没忍住.我对不起她.这种感觉以前从来没有过.我很难受,很难受.如果我能替她承担一些难过就好了.我放弃了一个这么好的女孩,我从来没遇到过的....都是我的错!我会后悔的,我今后也许也不会遇到象她这样无私的女孩了.她真的没有错,真的没有一点点错.希望她以后一切都好.一切都好. http://watershed.itpub.net/post/365/114188 http://watershed.itpub.net/post/365/114188 http://watershed.itpub.net/post/365/114188 watershed bbs Wed,14 06 2006 14:01:09 zhangsy之家 我终于败下阵了,对不起宝贝,快一年多的相处让我已经很习惯跟你在一起了.可是我想我该离开了,看到这里你可能很吃惊,可是我没有办法,这段时间我一直很矛盾,原因我想你是知道的,跟他们谈了好多次,最近一次就是昨天晚上,球赛也没看,一直在说这个事情,可是没有结果,我改变不了他们的看法.我想他们是为我好,但是这对你很不公平.跟你在一起的日子里我知道我有时候不够细心,让你受委屈了,也许象你说的那样:我有点太独了,对不起,请你原谅我.你可以恨我,但是你不能怀疑我跟你在一起的日子(尽管最近你可能有点不信任我,这让我有点生气).也许这对你也... http://watershed.itpub.net/post/365/113941 http://watershed.itpub.net/post/365/113941 http://watershed.itpub.net/post/365/113941 watershed bbs Tue,13 06 2006 14:31:43 zhangsy之家 正在不断更新中......1>10grman在backup的时候将当前没有用的到block跳过,不备份.而在以前的版本中rman只是跳过never been allocated blocks.2>The CONFIGURE DEVICE TYPE ... PARALLELISM command specifies the number ofautomatic channels to allocate for a specified device type. For example, if you configure parallelism to 3 for a device type, then RMAN allocates three channels for the device type when using automatic channels.3>并行通道的个数应该和物理设备的match,也就是说如果只有一个磁带设备,那么sbt通道就不... http://watershed.itpub.net/post/365/103794 http://watershed.itpub.net/post/365/103794 http://watershed.itpub.net/post/365/103794 watershed Tech Documents Fri,26 05 2006 15:32:40 zhangsy之家 TOAD要运行执行计划，必须先运行安装文件路径里的temptoadprep.sql http://watershed.itpub.net/post/365/60765 http://watershed.itpub.net/post/365/60765 http://watershed.itpub.net/post/365/60765 watershed Tech Documents Sat,18 03 2006 23:53:04 zhangsy之家