Kbase P19108: Database will not start with error (612), SYSTEM ERROR: Possible file truncation, <block-offset> too
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
3/19/2009 |
|
Status: Unverified
SYMPTOM(s):
Database will not start with error (612)
SYSTEM ERROR: Possible file truncation, <block-offset> too big for database. (612)
Corrupt block detected when attempting to release buffer.(4232)
Proutil dbname -C truncate bi fails with same errors.
Running proutil -C DBRPR to scan database.
Proutil dbname -C dbrpr with F option.
DBRPR reports multiple bad fragments in database.
CAUSE:
When a record spans two database blocks, it has the following structure RECID(1) and a pointer within RECID(1) to RECID(2).
Record fragments occur when the continuation pointer in RECID(1) to RECID(2) becomes corrupt negating RECID(1) from finding RECID (2). Or
if the Record id listed in RECID(1) as the continuation record no longer exists. In this situation RECID(1) is just a fragment because the rest of the record is not found.
FIX:
It is strongly suggested that a good backup of the database be restored. If this is not possible the database should be backed up in its current state.
If backup is not an alternative.
1. Run DBRPR to delete the bad fragments and their index entries.
syntax> proutil dbname -C DBRPR
2. From the Database repair menu select option 3, "Delete bad fragments".
3. Manual enter in the record fragments from the output of the scan.
4. After all record fragments have been deleted rebuild the index.
syntax> proutil dbname -C idxbuild all.