解決oracle 11g庫shutdown導致10g庫的crsd進程重啟辦法

客戶有個環境是10g的RAC，由於一次偶爾的需求，需要將一個11g的資料庫臨時在上面啟動，當我們mount了11g的軟體卷和datafile 卷之後，11g的資料庫能正常啟動，但是當11g的資料庫shutdown時，導致了10g的crsd進程重啟。

在10g的crsd的log中，可以看到：

2011-10-16 11:35:35.827: [ OCRAPI][267]procr_open_key: Invalid keyname [CRS.CUR.]. Component too big.
2011-10-16 11:35:35.827: [ CRSOCR][267] OCR api procr_open_key failed for key CRS.CUR.. OCR error code = 8 OCR error msg: PROC-8: Cannot perform cluster registry operation because one of the parameters is invalid.
2011-10-16 11:35:35.828: [ CRSOCR][267][PANIC] Failed to open key: CRS.CUR.(File: caaocr.cpp, line: 358)
2011-10-16 11:35:35.828: [ CRSD][267][PANIC] CRSD Exiting. OCR Failed
2011-10-16 11:35:35.828: [ CRSD][267] Done.

問題不是很大，2分鐘後，crsd自動重啟，恢複正常。期間資料庫對外的服務不受影響。

原因：

This is caused by Bug:13263017 - 11GR2 DATABASE SHUTDOWN CAUSING THE 10G CRS TO PANIC
 
From 11.2 onwards, database is designed to talk to local CRSD to notify its state changes. This change is incompatible with 10g CRSD causing the 10g CRSD to panic.

解決方案：

The bug has not been fixed yet, but the workaround is available:
Set "_notify_crs=FALSE" in pfile or spfile for 11gR2 database and restart the instance.
 
This should avoid pre-11gR2 CRSD panic.