联系：手机/微信(+86 13429648788) QQ(107644445)
作者：惜分飞©版权所有[未经本人同意,不得以任何形式转载,否则有进一步追究法律责任的权利.]
近期到现场进行了一个数据库恢复,我在恢复之前该库先由于硬件进行恢复,然后由其他人对其进行了一系列数据库恢复,但是未恢复成功,客户希望我们到现场进行处理(因为网络原因无法远程).接手库之后,处理第一个问题,是客户在进行现场备份的时候(把linux数据拷贝到win的过程中)发现有几个文件拷贝异常,这个错误很可能是由于当初的硬件故障修复之后留下的后遗症(由于io设备错误,无法运行此项请求),通过工具进行拷贝,恢复出来
DUL> copy file from /oradata2/xifenfeidata.dbf to /oradata2/xifenfeidata.dbfstarting copy datafile '/oradata1/xifenfeidata.dbf' to '/oradata2/xifenfeidata.dbf'read data error from file '/oradata1/xifenfeidata.dbf'.error message:Input/output errorread block# error: 560171read data error from file '/oradata1/xifenfeidata.dbf'.error message:Input/output errorread block# error: 560179datafile copy completed with 2 block error. |
[oracle@localhost ~]$ dbv file=/oradata2/xifenfeidata.dbf blocksize=16384DBVERIFY: Release 11.2.0.3.0 - Production on Mon Mar 29 17:28:17 2021Copyright (c) 1982, 2011, Oracle and/or its affiliates. All rights reserved.DBVERIFY - Verification starting : FILE = /oradata2/xifenfeidata.dbfPage 560171 is marked corruptCorrupt block relative dba: 0x3bc88c2b (file 239, block 560171)Completely zero block found during dbv: Page 560179 is marked corruptCorrupt block relative dba: 0x3bc88c33 (file 239, block 560179)Completely zero block found during dbv: DBVERIFY - Verification completeTotal Pages Examined : 4194302Total Pages Processed (Data) : 2230726Total Pages Failing (Data) : 0Total Pages Processed (Index): 1936953Total Pages Failing (Index): 0Total Pages Processed (Other): 26618Total Pages Processed (Seg) : 0Total Pages Failing (Seg) : 0Total Pages Empty : 3Total Pages Marked Corrupt : 2Total Pages Influx : 0Total Pages Encrypted : 0Highest block SCN : 304929867 (106.304929867) |
修复完相关无法拷贝文件之后,启动数据库报控制文件异常
Mon Mar 29 15:03:38 2021alter database mountUSER (ospid: 29044): terminating the instanceMon Mar 29 15:03:42 2021System state dump requested by (instance=1, osid=29044), summary=[abnormal instance termination].System State dumped to trace file /u01/app/oracle/diag/rdbms/xff/xff/trace/xff_diag_28961.trcInstance terminated by USER, pid = 29044 |
尝试重建ctl
[oracle@localhost ~]$ sqlplus / as sysdbaSQL*Plus: Release 11.2.0.3.0 Production on Mon Mar 29 17:40:17 2021Copyright (c) 1982, 2011, Oracle. All rights reserved.Connected to an idle instance.SQL> startup nomount pfile='/tmp/pfile'ORACLE instance started.Total System Global Area 1.7704E+10 bytesFixed Size 2235568 bytesVariable Size 2348811088 bytesDatabase Buffers 1.5301E+10 bytesRedo Buffers 52580352 bytesSQL> @/tmp/ctl.sqlCREATE CONTROLFILE REUSE DATABASE xff NORESETLOGS NOARCHIVELOG*ERROR at line 1:ORA-01503: CREATE CONTROLFILE failedORA-01189: file is from a different RESETLOGS than previous filesORA-01110: data file 249: '/oradata/xff/system03.dbf' |
初步判断是由于对方之前恢复导致部分文件resetlogs scn异常,通过bbed进行判断确认
BBED> set file 1 FILE# 1BBED> p kcvfhrlsstruct kcvfhrls, 8 bytes @116 ub4 kscnbas @116 0x00000001 ub2 kscnwrp @120 0x0000BBED> set file 249 FILE# 249BBED> p kcvfhrlsstruct kcvfhrls, 8 bytes @116 ub4 kscnbas @116 0x00000001 ub2 kscnwrp @120 0x0000 |
通过bbed修改相关值,然后重建控制文件成功,尝试resetlogs库，报ORA-01248错误
SQL> alter database open resetlogs;alter database open resetlogs*ERROR at line 1:ORA-01248: file 234 was created in the future of incomplete recoveryORA-01110: data file 234: '/oradata1/xifenfeidata5.DBF' |
关于ORA-01248的错误解释
01248, 00000, "file %s was created in the future of incomplete recovery"// *Cause: Attempting to do a RESETLOGS open with a file entry in the// control file that was originally created after the UNTIL time// of the incomplete recovery.// Allowing such an entry may hide the version of the file that // is needed at this time. The file number may be in use for// a different file which would be lost if the RESETLOGS was allowed.// *Action: If more recovery is desired then apply redo until the creation// time of the file is reached. If the file is not wanted and the// same file number is not in use at the stop time of the recovery,// then the file can be taken offline with the FOR DROP option.// Otherwise a different control file is needed to allow the RESETLOGS.// Another backup can be restored and recovered, or a control file can// be created via CREATE CONTROLFILE. |
大概的意思是文件的创建时间大于文件当前的scn,通过查询确实如此
SQL> select file#,CREATION_CHANGE#,CREATION_TIME from v$datafile_header where file#=234; FILE# CREATION_CHANGE# CREATION_---------------- ---------------- --------- 234 419298664864 02-AUG-19SQL> SELECT status, 2 to_char(checkpoint_change#,'9999999999999999') "SCN", 3 to_char(checkpoint_time,'yyyy-mm-dd hh24:mi:ss') checkpoint_time,FUZZY, 4 count(*) ROW_NUM 5 FROM v$datafile_header 6 GROUP BY status, checkpoint_change#, to_char(checkpoint_time,'yyyy-mm-dd hh24:mi:ss'),fuzzy 7 ORDER BY status, checkpoint_change#, checkpoint_time;STATUS SCN CHECKPOINT_TIME FUZ ROW_NUM------- ----------------- ------------------- --- ----------------ONLINE 417750848223 2021-02-23 23:50:46 YES 7ONLINE 417750848223 2021-03-21 11:44:25 NO 396 |
通过对部分scn进行修改（比如减小创建时间的scn)，然后尝试resetlogs库
SQL> alter database open resetlogs;alter database open resetlogs*ERROR at line 1:ORA-01092: ORACLE instance terminated. Disconnection forcedORA-00704: bootstrap process failureORA-00704: bootstrap process failureORA-00604: error occurred at recursive SQL level 1ORA-01555: snapshot too old: rollback segment number 5 with name"_SYSSMU5_2708889888$" too smallProcess ID: 3182Session ID: 1 Serial number: 3 |
这个错误比较简单,参考以前的部分文章:在数据库open过程中常遇到ORA-01555汇总数据库open过程遭遇ORA-1555对应sql语句补充,处理之后,数据库open成功
SQL> startup mount;ORACLE instance started.Total System Global Area 1.7704E+10 bytesFixed Size 2235568 bytesVariable Size 2348811088 bytesDatabase Buffers 1.5301E+10 bytesRedo Buffers 52580352 bytesDatabase mounted.SQL> alter database open;Database altered. |
本次数据库恢复基本上完成,已经最大限度恢复数据,导出数据到新库,完成恢复任务