Enterprise Manager

Hi,

I am testing EM10g Metrics at present. I have corrupted a data block and when I try to select it via a SQL*Plus session, it returns:

ERROR at line 1:
ORA-01578: ORACLE data block corrupted (file # 3, block # 1162)
ORA-01110: data file 3: '/ybs/u664/oradata/em9i/em9itestlocal00101.dbf'

However, in the alert log all that gets written is:

***
Corrupt block relative dba: 0x00c0048a (file 3, block 1162)
Bad header found during buffer read
Data in bad block -
type: 16 format: 2 rdba: 0x00c00489
last change scn: 0x0000.00012f48 seq: 0x3 flg: 0x04
consistency value in tail: 0x2f481003
check value in block header: 0x4c7e, computed block checksum: 0x0
spare1: 0x0, spare2: 0x0, spare3: 0x0
***

I suspect that the way in which the "Data Block Corruption Alert Log Error" metric works is to search the alert.log for the ORA-01578 error. But as this does not get written to the alert.log then it will never fire!!!?

Sure enough there is nothing in the alerts in the Grid Control console about the block corruption, and hence no e-mail alert.

Has anyone else come across this? Is this a bug with EM10g? Do I need to set an init.ora parameter to have block corruption errors written to the alert.log in a different way?

Any help appreciated.
Thanks in advance,
Steve.