oracle 12C上Error 16063的解药

**导读**

> 作者：杨漆

> 16年关系型数据库管理，从oracle 9i 、10g、11g、12c到Mysql5.5、5.6、5.7、8.0 到TiDB获得3个OCP、2个OCM；运维路上不平坦，跌过不少坑、熬过许多夜。把工作笔记整理出来分享给大伙儿，希望帮到大家少走弯路、少熬夜。

Rac主库日志中报出入下Error

TT02: Standby redo logfile selected for thread 1 sequence 289676 for destination LOG_ARCHIVE_DEST_4

TT02: Standby redo logfile selected for thread 1 sequence 289676 for destination LOG_ARCHIVE_DEST_3

TT02: Standby redo logfile selected for thread 1 sequence 289676 for destination LOG_ARCHIVE_DEST_2

2021-07-05T18:20:42.929962+08:00

TT02: Error 16063 archiving log 32 to 'orcldg3'

TT02: Error 16063 archiving log 32 to 'orclstb'

2021-07-05T18:20:44.167380+08:00

百度的解决方案:

ORA-16063: remote archival is enabled by another instanceCause: Remote archival of database REDO log files has been enabled by another instance, and cannot be disabled for this instance.Action: Set the REMOTE_ARCHIVE_ENABLE parameter to TRUE and restart this instance.

很开心，以为自己如此神速找到解药

在Master上执行:

SQL> alter system set remote_archive_enable='TRUE';

alter system set remote_archive_enable='TRUE'

*

ERROR at line 1:

ORA-25138: REMOTE_ARCHIVE_ENABLE initialization parameter has been made

obsolete

此参数在12C上已废弃

[捂脸]

搜遍全网 没找到解药。

技术圈里向大神请教，终于找到答案:

CAUSE

- Error 16063 is just informative that the destination is detached from the multi-async process and this happens only in environments that have multiple ASYNC destinations - The detachment happens because one of the destinations is slower than the other destinations. So it can not be collectively served by one process.

- In some situations the switching between single and multi async may be exacerbating the transport lag but the eviction indicates that a network issue is present before switching to a single async process.

SOLUTION

Solve the root cause and make sure all ASYNC destinations perform in similar way fixes the error ORA-16036 . If the error 16063 is one time, it may be temporary network issue with one of standbys that may be fixed. If customer have far standby , they can use farsync or cascaded standby also to fix this behavior.

Workaround:

Set the below parameter on primary and restart.

SQL> alter system set "_REDO_TRANSPORT_ASYNC_MODE"=1 scope=spfile sid='*'; --> this parameter needs database restart.

The parameter "_REDO_TRANSPORT_ASYNC_MODE"=1 disables this feature and we should be using it to assess the effect of the feature and compare lag happening when it is enabled versus when it is disabled.

故障解决！

在此特别感谢茫茫黑夜里给予我帮助和支持的小伙伴们！

把这个很难找到正确解药的方子分享出来，希望可以帮到在茫茫黑夜里同样艰难前行，排错到深夜的朋友，希望大家能少加班、少熬夜...