处理backup job 失败一例
今天查看运行在windows上的production数据库,发现上个星期五的备份job失败.报错误:ORA-27072: File I/O errorOSD-04008: WriteFile() failure, unable to write to fileO/S-Error: (OS 33) The process cannot access the file because another process has locked a portion of the file. 查看metalink,发现原来是防病毒软件的问题,解决步骤如下:Include the following OS commands before running the RMAN backup script:net stop "SavRoam"net stop "Symantec AntiVirus"net stop "Symantec AntiVirus Defin
Explain plan for and autotrace diffrence
The main difference have three points:1. Explain plan for : generate the execution plan not the actually execution plan will happen on database. the actually plan should query v$sql_plan view. Autotrace: generate the actually execution plan2. Explain plan for: only work at query statement that is SELECTs; Autotrace: works for any SQL and DML statements3.Explain plan for: you have to use utlxpls.sql to have the readable report for the execution plan.Autotrace: automaticaly show the execution plan
understanding oracle 10g incremental updated backup using rman
those days i try to understand incremental updated backup, example :RMAN> run {2> recover copy of database with tag 'level_0';3> backup incremental level 1 for recover of copy with tag 'level_0' database format '/backup/%UORA10G';4> }Starting recover at 02-JAN-08using channel ORA_DISK_1no copy of datafile 1 found to recoverno copy of datafile 2 found to recoverno copy of datafile 3 found to recoverno copy of datafile 4 found to recoverFinished recover at 02-JAN-08Starting backup at 0
transport database between diffrent platform
transport datase between diffrent os platform linux to windos1.open source database song on linux platform read only:SQL> startup mountORACLE instance started.Total System Global Area 494927872 bytesFixed Size 1220192 bytesVariable Size 167772576 bytesDatabase Buffers 322961408 bytesRedo Buffers 2973696 bytesDatabase mounted.SQL> alter database open read only;Database altered.2.support platform list:source database platform information:
Oracle在windows下有关注册表项
1.HKEY_LOCAL_MACHINESOFTWAREORACLEHOMEID:includes the following parameters for an Oracle home directory.2.HKEY_LOCAL_MACHINESOFTWAREORACLE:3.HKEY_LOCAL_MACHINESOFTWAREORACLEALL_HOMES:This subkey provides general information on each Oracle home directory on a computer.4.HKEY_LOCAL_MACHINESYSTEMCurrentControlSetServices: display those services related to oracle5.HKEY_LOCAL_MACHINESOFTWAREORACLEOSD9I:RAC related registry item.6.HKEY_LOCAL_MACHINESOFTWAREORACLEOSD9ICM:This subkey contains Cluster Ma
关于口令文件的一些总结
关于口令文件的我们有这样或者那样的误解,在此做一个总结。。[@more@]口令文件的我们可以从参数文件中设置其作用,remote_login_passwordfile参数。这个文件的作用是验证具有sysdba and sysoper权限的数据库用户,对应口令保存在这个文件中,当然是加密了的口令了(不然我们打开这个文件就可以随意使用sysdba登录数据库了)对于具有三个选项:NONE:忽略口令文件的存在,远程登录数据库时需要使用os验证。EXCLUSIVE:在远程登录数据库时使用口令文件进行验证,并且这个口令文件只能被一个数据库使用。SHARED:口令文件可以被多个数据库使用
关于两个性能相关参数的设置
optimizer_index_caching & optimizer_index_cost_adj[@more@]怎么样设置这两个参数呢?有个通用的指导建议:系统特征为DSS的:optimizer_index_caching=0 ,optimizer_index_cost_adj=100.(Default set)系统特征为OLTP的:optimizer_index_caching=80~90 ,optimizer_index_cost_adj=25~50这两个参数只是会影响到执行计划的选择,并不会影响到性能,所以我理解这两个参数只是会给系统一个选择不同执行计划的机会。如果系统用10g的话,不建议改变其缺省设置。通常我们要定期收集system statistcs就是我们可以通过准确的system statistic
做系统跟踪的几个备忘录
关于在OS and database level 执行跟踪。。。[@more@]kenel trace: 1.check os log 2. tools: truss :aix . solaris ,hp strace: linuxnetwork trace: 1.os level : netstat -an, lsof, snoop 2.database level : client side sqlnet.log ,tnsping ,lsnrctl oracle event trace : 1.specific error :set 6502 event(alter session ,dbms_system.set_ev.2.performance issuses: set 10046 event .3.trace file location : user_dump_dest.
How to troubleshoot oracle error my checklist!
My checklist for troubleshoot oracle RDBMS error...[@more@]step 1: Check Os system log . sun solaris : /var/adm/messages. aix: errpt command. hp-ux: /var/adm/syslog/syslog.log. linux : /var/log/messages. windows : event viewer.step2: check oracle alertlog and tracefile.step3: Metalink, oracle support, forum and google.com
关于oracle logical and physical block corruption
主要的检测方法有以下几种:1.exp/imp2.analyze table xxx validate structure cascade3.dbv4.可以设置db_block_checking -->memory errordb_block_checksum --->disk error5.rman validate command 6.OS command: sum[@more@]关于
数据库备份
Database hotbackup:1.rman 2.alter tablespace begin/end backup(require database running on archvielog mode).Database coldbackup:1.rman(mount mode).2.cp/copy paste.including datafile,controlfile,online redo log file.parameter file ,password file, network associated files.Archived redolog files:1.rman 2.after database hotbackup ,copy every archived redolog file.Controlfile:1.rman(using sql command to backup controlfile finally) 2.alter database backup controlfile to trace/ /derictory/filename.Para
备份操作系统
几个工具:solaris:dump,tar,cpio,dd. AIX: smitty mksysbHp:fbackup,sam.Windows :ghost...[@more@]