数据库突然宕机无法open的问题及解决

测试的数据库有一天突然宕机，然后无法正常open了，这个问题虽然过去了一段时间，也在这儿总结一把。 从alert日志中的信息如下。

Fri Jan 10 16:09:42 2014
Archived Log entry 6837 added for thread 1 sequence 6863 ID 0x19db56aa dest 1:
Fri Jan 10 16:12:59 2014
Errors in file /oravl01/oracle/adm/FETABP2/diag/rdbms/fetabp2/FETABP2/trace/FETABP2_lgwr_2432.trc:
ORA-19502: write error on file "/oravl04/oradata/FETABP2/cntrl_1.dbf", block number 1191 (block size=16384)
ORA-27061: waiting for async I/Os failed
Linux-x86_64 Error: 28: No space left on device
Additional information: -1
Additional information: 131072
LGWR (ospid: 2432): terminating the instance due to error 19502
Fri Jan 10 16:13:01 2014
System state dump requested by (instance=1, osid=2432 (LGWR)), summary=[abnormal instance termination].
System State dumped to trace file /oravl01/oracle/adm/FETABP2/diag/rdbms/fetabp2/FETABP2/trace/FETABP2_diag_2422.trc
Non critical error ORA-48913 caught while writing to trace file "/oravl01/oracle/adm/FETABP2/diag/rdbms/fetabp2/FETABP2/trace/FETABP2_diag_2422.trc"
Error message: ORA-48913: Writing into trace file failed, file size limit [5242880] reached
Writing to the above trace file is disabled for now on...
Instance terminated by LGWR, pid = 2432
Fri Jan 10 16:32:32 2014

从上可以看到，数据库遇到了io问题，并且空间也不够了，直接宕机了。 先mount上再说，别直接拿过来就open，可能一些恢复问题让自己的误操作弄的更复杂了。如果生产环境，那影响就更大了。需要先做详细的判断再动手。 由于这个是测试环境先来演示一下错误。

alter database open
Errors in file /oravl01/oracle/adm/FETABP2/diag/rdbms/fetabp2/FETABP2/trace/FETABP2_ora_4780.trc:
ORA-10873: file 1 needs to be either taken out of backup mode or media recovered
ORA-01110: data file 1: '/oravl04/oradata/FETABP2/SYSTEM_1.dbf'
ORA-10873 signalled during: alter database open...
Fri Jan 10 17:02:26 2014

查看数据库状态。 SQL> select status from v$instance; STATUS ------------ MOUNTED 查看数据文件的scn情况 SQL> select file#,checkpoint_change# from v$datafile; FILE# CHECKPOINT_CHANGE# ---------- ------------------ 1 1.3605E+13 2 1.3605E+13 3 1.3605E+13 4 1.3605E+13 5 1.3605E+13 6 1.3605E+13 7 1.3605E+13 8 1.3605E+13 9 1.3605E+13 10 1.3605E+13 11 1.3605E+13 11 rows selected. 显示不清楚，先格式化再看看。 SQL> col checkpoint_change# format 99999999999999999 SQL> / FILE# CHECKPOINT_CHANGE# ---------- ------------------ 1 13605259062393 2 13605259062399 3 13605259062404 4 13605259062411 5 13605259062416 6 13605259062422 7 13605259062411 8 13605259062411 9 13605259062411 10 13605259062411 11 13605259062411 11 rows selected. 可以看到有很多不一致。 查看数据库文件头的scn情况，情况类似。 SQL> select file#,checkpoint_change# from v$datafile_header; FILE# CHECKPOINT_CHANGE# ---------- ------------------ 1 13605259062393 2 13605259062399 3 13605259062404 4 13605259062411 5 13605259062416 6 13605259062422 7 13605259062411 8 13605259062411 9 13605259062411 10 13605259062411 11 13605259062411 如果是这个状态，可能是有对数据库做了什么其他的操作,查看热备份的情况 这下揪到问题了。 SQL> select * from v$backup; FILE# STATUS CHANGE# TIME ---------- ------------------ ---------- --------- 1 ACTIVE 1.3605E+13 08-JAN-14 2 ACTIVE 1.3605E+13 08-JAN-14 3 ACTIVE 1.3605E+13 08-JAN-14 4 ACTIVE 1.3605E+13 08-JAN-14 5 ACTIVE 1.3605E+13 08-JAN-14 6 ACTIVE 1.3605E+13 08-JAN-14 7 ACTIVE 1.3605E+13 08-JAN-14 8 ACTIVE 1.3605E+13 08-JAN-14 9 ACTIVE 1.3605E+13 08-JAN-14 10 ACTIVE 1.3605E+13 08-JAN-14 11 ACTIVE 1.3605E+13 08-JAN-14 从日志里面翻看热备份的情况，连续好几天都没有end backup的命令出现，可见io的那个问题和这个也有一定的关系。 先修复一下。 用下面的sql生成修复语句。 select 'alter tablespace '||name|| ' end backup;' from v$tablespace where ts# in (select ts# from v$datafile where file# in (select file# from v$backup));

生成的命令如下。 alter tablespace system end backup; ..... 修复了一部分，查看备份表。可以看到好多状态都发生了改变。 SQL> select * from v$backup; FILE# STATUS CHANGE# TIME ---------- ------------------ ---------- --------- 1 NOT ACTIVE 1.3605E+13 08-JAN-14 2 NOT ACTIVE 1.3605E+13 08-JAN-14 3 ACTIVE 1.3605E+13 08-JAN-14 4 NOT ACTIVE 1.3605E+13 08-JAN-14 5 NOT ACTIVE 1.3605E+13 08-JAN-14 6 NOT ACTIVE 1.3605E+13 08-JAN-14 7 NOT ACTIVE 1.3605E+13 08-JAN-14 8 NOT ACTIVE 1.3605E+13 08-JAN-14 9 NOT ACTIVE 1.3605E+13 08-JAN-14 10 NOT ACTIVE 1.3605E+13 08-JAN-14 11 NOT ACTIVE 1.3605E+13 08-JAN-14 11 rows selected. 修复完成后，可以看到状态都是not active的了。 SQL> select * from v$backup; FILE# STATUS CHANGE# TIME ---------- ------------------ ---------- --------- 1 NOT ACTIVE 1.3605E+13 08-JAN-14 2 NOT ACTIVE 1.3605E+13 08-JAN-14 3 NOT ACTIVE 1.3605E+13 08-JAN-14 4 NOT ACTIVE 1.3605E+13 08-JAN-14 5 NOT ACTIVE 1.3605E+13 08-JAN-14 6 NOT ACTIVE 1.3605E+13 08-JAN-14 7 NOT ACTIVE 1.3605E+13 08-JAN-14 8 NOT ACTIVE 1.3605E+13 08-JAN-14 9 NOT ACTIVE 1.3605E+13 08-JAN-14 10 NOT ACTIVE 1.3605E+13 08-JAN-14 11 NOT ACTIVE 1.3605E+13 08-JAN-14 再次查看scn的情况。 SQL> select file#,checkpoint_change# from v$datafile 2 SQL> / FILE# CHECKPOINT_CHANGE# ---------- ------------------ 1 13607479649688 2 13607479649688 3 13607479649688 4 13607479649688 5 13607479649688 6 13607479649688 7 13607479649688 8 13607479649688 9 13607479649688 10 13607479649688 11 13607479649688 查看数据文件头的情况。 SQL> select file#,checkpoint_change# from v$datafile_header; FILE# CHECKPOINT_CHANGE# ---------- ------------------ 1 13607479649688 2 13607479649688 3 13607479649688 4 13607479649688 5 13607479649688 6 13607479649688 7 13607479649688 8 13607479649688 9 13607479649688 10 13607479649688 11 13607479649688 确认没问题了，打开数据库。 SQL> alter database open; Database altered.