基于ORA-19815闪回空间爆满问题的处理方法

闪回区爆满问题也是经常会遇到的问题，最关键的是闪回设置大小以及归档被默认存放在了闪回目录，恰巧今天又遇到了这个问题，就记录下处理步骤，仅供遇到这类问题的人参考。

一、错误现象描述

1）应用端错误信息

Error: 2016-11-26 11:45:25 ORA-00257: Archiver error. Connect AS SYSDBA only until resolved.
Error: 2016-11-26 11:45:25 ORA-00257: Archiver error. Connect AS SYSDBA only until resolved.
Error: 2016-11-26 11:45:25 init connpool:one or more conn open error.

2）数据库端错误信息

Sat Nov 26 12:13:14 2016
Errors in file /home/U01/app/oracle/diag/rdbms/testdb/testdb/trace/testdb_arc2_929.trc:
ORA-19815: WARNING: db_recovery_file_dest_size of 42949672960 bytes is 100.00％ used, and has 0 remaining bytes available.
Sat Nov 26 12:13:14 2016
************************************************************************
You have following choices to free up space from recovery area:
1. Consider changing RMAN RETENTION POLICY. If you are using Data Guard,
 then consider changing RMAN ARCHIVELOG DELETION POLICY.
2. Back up files to tertiary device such as tape using RMAN
 BACKUP RECOVERY AREA command.
3. Add disk space and increase db_recovery_file_dest_size parameter to
 reflect the new space.
4. Delete unnecessary files using RMAN DELETE command. If an operating
 system command was used to delete files, then use RMAN CROSSCHECK and
 DELETE EXPIRED commands.
************************************************************************
Sat Nov 26 12:13:14 2016
Errors in file /home/U01/app/oracle/diag/rdbms/testdb/testdb/trace/testdb_arc2_929.trc:
ORA-19809: limit exceeded for recovery files
ORA-19804: cannot reclaim 524288000 bytes disk space from 42949672960 limit

二、错误分析

从应用日志看，是由于不能归档导致的DB连接池不能被初始化打开，基本可以判断是由于数据库的归档文件所在磁盘空间满导致。

从数据库日志查看发现确实是由于闪回空间已经爆满，通过查看服务器磁盘空间以及闪回空间大小即可再次验证。

三、详细处理过程

1）登陆数据库服务器，查看磁盘空间使用信息

[oracle@teststd trace]$ df -h
Filesystem   Size Used Avail Use％ Mounted on
/dev/sda5    9.9G 2.6G 6.9G 28％ /
tmpfs      32G  18M  32G  1％ /dev/shm
/dev/sda1    388M  62M 307M 17％ /boot
/dev/sda6    1.6T 506G 1017G 34％ /home
/dev/sda2    20G 508M  19G  3％ /var

2）查看数据库错误日志，发现闪回空间爆满

cd /home/U01/app/oracle/diag/rdbms/testdb/testdb/trace
tail -n 35 alert_testdb.log

ORA-19815: WARNING: db_recovery_file_dest_size of 42949672960 bytes is 100.00％ used, and has 0 remaining bytes available.

3）登陆数据库，查看闪回路径以及闪回空间使用情况

查看闪回空间设置大小

SQL> show parameter recover;

NAME         TYPE    VALUE
------------------------------------ ----------- ------------------------------
db_recovery_file_dest   string   /home/U01/app/oracle/fast_recovery_area
db_recovery_file_dest_size      big integer 40G
db_unrecoverable_scn_tracking    boolean   TRUE
recovery_parallelism         integer   0

或者利用SQL查看闪回路径

Col name for a60;
Set line 200;
SQL> select * from v$recovery_file_dest ;

NAME                  SPACE_LIMIT SPACE_USED SPACE_RECLAIMABLE NUMBER_OF_FILES   CON_ID
------------------------------------------------------- ----------- ---------- ----------------- --------------- ----------
/home/U01/app/oracle/fast_recovery_area         4.2950E+10 2.1538E+10         0       89     0

查看闪回空间所在磁盘大小

SQL> !df -h /home/U01/app/oracle/fast_recovery_area
Filesystem   Size Used Avail Use％ Mounted on
/dev/sda6    1.6T 504G 1019G 34％ /home

查看闪回目录使用大小

SQL> !du -hs /home/U01/app/oracle/fast_recovery_area
40G /home/U01/app/oracle/fast_recovery_area

查看闪回空间使用情况

SQL> select * from V$RECOVERY_AREA_USAGE;

FILE_TYPE    PERCENT_SPACE_USED PERCENT_SPACE_RECLAIMABLE NUMBER_OF_FILES   CON_ID
----------------------- ------------------ ------------------------- --------------- ----------
CONTROL FILE               0             0        0     0
REDO LOG           7.45             0        7     0
ARCHIVED LOG             100             0       76     0
BACKUP PIECE               0             0        0     0
IMAGE COPY                0             0        0     0
FLASHBACK LOG              0             0        0     0
FOREIGN ARCHIVED LOG           0             0        0     0
AUXILIARY DATAFILE COPY         0             0        0     0

4）删除归档、调整闪回区大小

从上面看确实是闪回空间满了，而占用闪回区满的罪魁祸首就是归档日志。要解决的方法有两个：

方案1:删除多余的归档

最佳删除归档的途径是通过rman工具做，如果直接删除文件数据库是识别不到闪回区释放的。

[oracle@teststd trace]$ rman target /
Recovery Manager: Release 12.1.0.2.0 - Production on Sat Nov 26 13:00:28 2016
Copyright (c) 1982, 2014, Oracle and/or its affiliates. All rights reserved.
connected to target database: TESTDB (DBID=2708971821)
RMAN> crosscheck archivelog all;
RMAN> delete expired archivelog all;
RMAN> DELETE NOPROMPT ARCHIVELOG ALL COMPLETED BEFORE 'SYSDATE-7' ;

或者利用下面的语句删除七天前的归档日志

RMAN> DELETE NOPROMPT ARCHIVELOG UNTIL TIME 'SYSDATE-7' ;

注意：（一般删除归档腾出足够的闪回区后，建议调整归档路径或者调整足够闪回区大小）

方案2：调整闪回区大小

SQL> alter system set db_recovery_file_dest_size=100G scope=both;
alter system set db_recovery_file_dest_size=100G scope=both
*
ERROR at line 1:
ORA-65040: operation not allowed from within a pluggable database
###数据库为12c，不允许在pdb下操作，需要切换到sys下操作
SQL> conn /as sysdba;
Connected.
SQL> alter system set db_recovery_file_dest_size=100G scope=both;

System altered.

5）查看闪回区使用情况和其大小

[oracle@teststd trace]$ du -hs /home/U01/app/oracle/fast_recovery_area
27G   /home/U01/app/oracle/fast_recovery_area

SQL> select * from V$RECOVERY_AREA_USAGE;
FILE_TYPE    PERCENT_SPACE_USED PERCENT_SPACE_RECLAIMABLE NUMBER_OF_FILES   CON_ID
----------------------- ------------------ ------------------------- --------------- ----------
CONTROL FILE           0       0         0     0
REDO LOG          2.98        0         7     0
ARCHIVED LOG         17.08       0         82     0
BACKUP PIECE           0       0         0     0
IMAGE COPY            0       0         0     0
FLASHBACK LOG          0       0         0     0
FOREIGN ARCHIVED LOG       0       0         0     0
AUXILIARY DATAFILE COPY         0     0        0     0

8 rows selected.

6）通知系统管理员启动应用，正常启动

至此问题得以彻底解决

来源：http://blog.itpub.net/27067062/viewspace-2129130/