節點2線上日誌生成歸檔日誌在節點1上的初步分析

    oracle rac 2節點

 

昨天去客戶現場做資料庫安全方面的調研，客戶問到一個問題，怎麼節點2的線上日誌生成的歸檔日誌怎麼跑到節點1上去了

 

分析過程大概如下:

節點2down機:
lert_gzgjj2.log (0%)
ORA-27301: OS failure message: Invalid argument
ORA-27302: failure occurred at: sskgpwpost0
ORA-27303: additional information: ukid = 874057518489215114, errno = 22
LMON: terminating instance due to error 29702
Sat Jul 30 13:26:00 2011
Errors in file /oracle/admin/gzgjj/bdump/gzgjj2_lms2_4540.trc:
ORA-29702: error occurred in Cluster Group Service operation
Sat Jul 30 13:26:00 2011
Errors in file /oracle/admin/gzgjj/bdump/gzgjj2_lms0_4536.trc:
ORA-29702: error occurred in Cluster Group Service operation
Sat Jul 30 13:26:00 2011
System state dump is made for local instance
System State dumped to trace file /oracle/admin/gzgjj/bdump/gzgjj2_diag_4528.trc
Sat Jul 30 13:26:04 2011
Instance terminated by LMON, pid = 4532

 

節點1 13：26分日誌如下：

Started redo application at
 Thread 2: logseq 9705, block 99429
Sat Jul 30 13:26:03 2011
Recovery of Online Redo Log: Thread 2 Group 6 Seq 9705 Reading mem 0
  Mem# 0: +DG1/gzgjj/onlinelog/group_6.268.651752891
Sat Jul 30 13:26:03 2011
Recovery of Online Redo Log: Thread 2 Group 5 Seq 9706 Reading mem 0
  Mem# 0: +DG1/gzgjj/onlinelog/group_5.267.651752889
Sat Jul 30 13:26:05 2011
Completed redo application
Sat Jul 30 13:26:06 2011
Completed instance recovery at
 Thread 2: logseq 9706, block 49566, scn 1041346872
 7450 data blocks read, 7479 data blocks written, 50484 redo blocks read
Switch log for thread 2 to sequence 9707
Sat Jul 30 13:26:09 2011
ARC1: Archiving disabled thread 2 sequence 9707

 

節點1做例項恢復，生成了節點2 的9706和9707兩個歸檔日誌

 

-rw-rw----   1 oracle     dba        50755584 Jul 30 13:26 2_9706_658773440.dbf
-rw-rw----   1 oracle     dba           2048 Jul 30 13:26 2_9707_658773440.dbf

 

分析總結如下:

 

   節點2在2011.07.30 13:26 分發生down機，而這個時候業務漂移到節點1上，節點1做資料庫例項恢復，該前滾的前滾，

該回滾的回滾，所以節點2的歸檔日誌產生到節點1上。