Re: Corrupt block message in alert log

Run the following:

select segment_name,segment_type,owner
from dba_extents
where file_id = <f> and <b> between block_id and block_id + blocks - 1;

--where <f> is the file number reported in the error message and <b> is the block number reported in the error message. to find exactly what object is corrupted. Once you have the segment, you should perform an ANALYZE on the object with the VALIDATE STRUCTURE CASCADE clause. This may or may not produce an error. If an error is generated, there should be a corresponding trace file in your user_dump_dest directory. This would likely signal a true corruption in the ojbect. If it is a data dicationary object, the only recourse is to either restore from backup or recreate the database (i.e. full export, recreate, full import). Depending on the object involved, there may be an issue with generating an export.

"Lynux" <lynux_at_netvigator.com> wrote in message news:93q7ia$l4135_at_imsp212.netvigator.com...
> Recently, we found the following message appearred in our alert_<SID>.log
> file.
>
>
> ***
> Corrupt block relative dba: 0x1e403ac6 file=121. blocknum=15046.
> Fractured block found during direct read
> Data in bad block - type:8. format:2. rdba:0x1e403ac8
> last change scn:0x0000.382b9618 seq:0x1 flg:0x08
> consistancy value in tail 0x96840801
> check value in block header: 0x0, check value not calculated
> spare1:0x0, spare2:0x0, spare2:0x0
> ***
> Corrupt block relative dba: 0x1e403ac8 file=121. blocknum=15048.
> Fractured block found during direct read
> Data in bad block - type:8. format:2. rdba:0x1e403ac8
> last change scn:0x0000.382b9618 seq:0x1 flg:0x08
> consistancy value in tail 0x96840801
> check value in block header: 0x0, check value not calculated
> spare1:0x0, spare2:0x0, spare2:0x0
>
>
> However, when we try to see any problem on the corresponding tablespace,
> the result is everything normal. The size reported in system view is the
> same as before. The 'v$recover_file' return no datafile need to be
> repaired. The 'alter system check datafiles;' return everything are
> normal.
>
> Do anyone know what's the problem on these messages. Please help and
> advice. Thank You Very Much!
> Thanks in Advance!
Received on Sat Jan 13 2001 - 18:38:19 CST