sqlnet trace

admin

一步一步地解释如何在数据库环境上动态生成SQLNET（服务器跟踪）跟踪。
1.主备库配置sqlnet.ora

1. TRACE_DIRECTORY_SERVER = /TMP
   
2. DIAG_ADR_ENABLED = OFF

复制代码

2.主备库开启archive log tracing
2.1清除/移动主目录和备用目录的现有跟踪转储目录。
2.2在主上，可选择推迟REDO传输。

1. SQL> alter system set log_archive_dest_state_2=defer;
   
2. SQL> alter system switch logfile;
   
3. SQL> alter system set log_archive_trace=4095;
   

复制代码

2.3重启arch进程

1. ps -ef|grep ora_arc|grep -v grep|awk '{print $2}'|xargs -i kill -9 {}

复制代码

3.在主节点上，重新启用目标以阻止重做传输以重新启动。

1. SQL> alter system set log_archive_dest_state_2=enable;
   
2. SQL> alter system switch logfile;

复制代码

5.在日志传输期间，监视待机状态上的v$托管_待机视图，以捕获ARCH/RFS关系。

1. SQL> select inst_id, process, pid, client_process, client_pid, status,thread#,sequence# from gv$managed_standby;

复制代码

6.对于动态SQLNET跟踪
6.1 a) For 11.1.0.7, 11.2.0.2, 11.2.0.3

1. On primary side,
   
2. oradebug setospid <client PID from the above query>
   
3. oradebug dump event_tsm_test 16
   
4. oradebug tracefile_name
   
5. 
   
6. To unset the tracing,
   
7. oradebug dump event_tsm_test 0
   
8. 
   
9. On standby side,
   
10. oradebug setospid <PID from the above query>
    
11. oradebug dump event_tsm_test 16
    
12. oradebug tracefile_name

复制代码

b) From 11.2.0.4 onwards,

1. On primary side,
   
2. 
   
3. oradebug setospid <client PID from the above query>
   
4. oradebug dump sqlnet_server_trace 16
   
5. oradebug tracefile_name
   
6. 
   
7. On standby side,
   
8. 
   
9. oradebug setospid <PID from the above query>
   
10. oradebug dump sqlnet_server_trace 16
    
11. oradebug tracefile_name
    
12. 
    
13. To unset the tracing on Primary and Standby,
    
14. 
    
15. oradebug dump sqlnet_server_trace 0

复制代码

7.一旦生成了错误，就应该停止跟踪。

1. Set LOG_ARCHIVE_TRACE=0
   
2. 
   
3. on both primary and standby
   
4. Find and kill the ARCn processes per #2
   
5. Re-enable dest and switch log per #3

复制代码