10g RAC fail over測試，節點2例項不能隨Linux自動啟動

Linux ecsrac01 2.6.9-42.ELsmp #1 SMP Linux   

Linux 4.0 u5  64bit  
Oracle 10g  RAC   (10.2.0.3)

開始安裝成功之後進行fail over測試，手動關閉rac01的vip對應的實體網路卡，節點1  vip 自動飄移
到節點2 上，恢復節點1的網路卡正常後， 自動飄回去也正常，兩個例項都可以訪問，都是開啟狀態。

但是，正常關閉crs, 資料庫，然後重新啟動，結果發現節點1  public 網路卡有兩個vip ,  節點2沒有vip,
節點1， 2 監聽都不能正常隨OS重新啟動而啟動， 節點1 例項啟動了，節點2 例項沒有啟動，一段時間
後， 節點1， 2 的vip都恢復正常(都有自己的vip) , 但是監聽還是都不行(沒有手動開啟)，節點2 的例項
也沒有開啟 。 crs_stat  -t  發現節點2的lis , ons , ins ... 都是unknown 狀態 。  檢視節點2的oracle
log , 發現沒有開啟資料庫的log .  



Tue Nov 18 09:56:40 2008
Thread 2 advanced to log sequence 50
  Current log# 12 seq# 50 mem# 0: /ocfs_data1/ecsdb/redo12a.log
  Current log# 12 seq# 50 mem# 1: /ocfs_index1/ecsdb/redo12b.log
Tue Nov 18 14:40:18 2008
Errors in file /u01/product/admin/ecsdb/bdump/ecsdb2_lmon_3536.trc:
ORA-29740: evicted by member 0, group incarnation 6
Tue Nov 18 14:40:18 2008
LMON: terminating instance due to error 29740
Tue Nov 18 14:40:18 2008
Errors in file /u01/product/admin/ecsdb/bdump/ecsdb2_lms0_3540.trc:
ORA-29740: evicted by member , group incarnation
Tue Nov 18 14:40:18 2008
Errors in file /u01/product/admin/ecsdb/bdump/ecsdb2_lms1_3544.trc:
ORA-29740: evicted by member , group incarnation
Tue Nov 18 14:40:18 2008
Errors in file /u01/product/admin/ecsdb/bdump/ecsdb2_lmd0_3538.trc:
ORA-29740: evicted by member , group incarnation
Tue Nov 18 14:40:18 2008
System state dump is made for local instance
System State dumped to trace file /u01/product/admin/ecsdb/bdump/ecsdb2_diag_3532.trc
Tue Nov 18 14:40:20 2008
Trace dumping is performing id=[cdmp_20081118144018]
Tue Nov 18 14:40:23 2008