[轉載] ALERT中頻繁出現DISKGROUP ARCH WAS DISOUNTED/MOUNTED

http://space6212.itpub.net/post/12157/516460

ALERT中頻繁出現DISKGROUP ARCH WAS DISOUNTED/MOUNTED

 

最近對ASM進行重新規劃後，在alert中頻繁發現以下資訊，很是煩人：


Thread 4 advanced to log sequence 3499 (LGWR switch)
Current log# 56 seq# 3499 mem# 0: +DBSYS/dwrac/redo01_56.log
Wed Apr 13 05:29:59 CST 2011
SUCCESS: diskgroup ARCH was mounted
Wed Apr 13 05:31:59 CST 2011
SUCCESS: diskgroup ARCH was dismounted
Wed Apr 13 05:51:56 CST 2011
Thread 4 advanced to log sequence 3500 (LGWR switch)
Current log# 57 seq# 3500 mem# 0: +DBSYS/dwrac/redo01_57.log
Wed Apr 13 05:51:56 CST 2011
SUCCESS: diskgroup ARCH was mounted
Wed Apr 13 05:52:20 CST 2011
SUCCESS: diskgroup ARCH was dismounted
Wed Apr 13 06:50:30 CST 2011
Thread 4 advanced to log sequence 3501 (LGWR switch)
Current log# 58 seq# 3501 mem# 0: +DBSYS/dwrac/redo01_58.log
Wed Apr 13 06:50:30 CST 2011
SUCCESS: diskgroup ARCH was mounted
Wed Apr 13 06:50:49 CST 2011
SUCCESS: diskgroup ARCH was dismounted
Wed Apr 13 08:42:03 CST 2011
Thread 4 advanced to log sequence 3502 (LGWR switch)
Current log# 59 seq# 3502 mem# 0: +DBSYS/dwrac/redo01_59.log


這是因為一個bug：4465015導致的。不過Oracle也說這是一個正常的行為：
當關閉某個disk group上的最後一個檔時，Oracle會將這個diskgroup dismount；
當需要讀取某個disk group某一個檔時，Oracle再把它mount。

在這個案例中，歸檔日誌放在一個獨立的disk group中，而歸檔日誌是隔段時間才會產生一個，所以會頻繁在alert日誌中看到這個資訊。
這個資訊是正常的，但是如果你覺得比較煩，可以把控制檔、redo的一個副本或者放一個很小的資料檔案到這個disk group上，依賴CKPT、LGWR、DWR幾個進程分別需要頻繁訪問以上檔來避免。

如：
SQL> alter tablespace admin add datafile '+ARCH/dwrac/datafile/admin02.dbf' size 10m;

Tablespace altered