| CORRUPT BLOCK, BAD HEADER FOUND DURING BUFFER READ [message #61720] |
Thu, 20 May 2004 11:21  |
DM
Messages: 3
Registered: April 2004
|
Junior Member |
|
|
Lately, my production databases have often reported Corruption issue which turns out to be a false alarm. The error that shows up in alert.log is usually:
***
Corrupt block relative dba: 0x04811440 (file 18, block 70720)
Bad header found during buffer read
Data in bad block -
type: 112 format: 0 rdba: 0x0000000a
last change scn: 0x4e00.00000000 seq: 0x1 flg: 0x00
consistency value in tail: 0xebcd0601
check value in block header: 0x0, block checksum disabled
spare1: 0xb, spare2: 0x0, spare3: 0x0
***
Reread of rdba: 0x04811440 (file 18, block 70720) found same corrupted data
If I restart the database, the error is gone. Until then the error repeats each time someone tried to access above block.
It has happened in 3 different databases on separate servers running 8.1.7.4 on Solaris 8. No errors in unix logs or SAN logs.
Any clue? Oracle did not see any corruption in blocks/files etc.
I want to turn on db_block_checking and db_block_checksum to true (OLTP db with 20% writes). Is there any problem with that in terms of corruption handling and load on the box?
I am planning to relink oracle just in case.
Any help would be appreciated.
|
|
|
|
|
|
] 
Blog
Search
Help
Members
Register
Login
Home
