Kbase P32484: DBDOWN with 9446 during write to bi file 3645
| Autor |
Progress Software Corporation - Progress |
| Acesso |
Público |
| Publicação |
5/6/2008 |
|
Status: Verified
FACT(s) (Environment):
All Supported Operating Systems
Progress/OpenEdge Versions
SYMPTOM(s):
DBDOWN with 9446 during write to bi file 3645
SYSTEM ERROR: bkioWrite: Bad file descriptor was used during write, fd 816, len 8192, offset 330, file d:\..\..\dbname.b1. (9446)
BROKER 0: bkwrite: write to disk failed errno 9. (3645)
BROKER 0: SYSTEM ERROR: Releasing regular latch. latchId: 1 (5028)
BROKER 0: User 0 died holding 1 shared memory locks. (2522)
BROKER 0: Begin ABNORMAL shutdown code 2 (2249)
BROKER : Multi-user session end. (334)
Veritas running but not against database file locations
Shared memory in use by another process. (1260)
When truncating the bi or accessing the database after crash errors 9446 during write to bi file 3645
CAUSE:
During write to the bi file, error 9: "The storage control block address is invalid." The shared memory pointer is corrupt.
FIX:
The shared memory pointer is corrupt, so the shared memory id (shmid) needs to be removed from the bi file and the master block and then bound to a new shmid instead of binding to the original shmid.
This can be done by either:
1. Rebooting the Server to clear shared memory.
2. prostrct repair on the database (to clean out shmid in the master block).
3. truncate the bi file to remove the shmid from the bi file. Then add the bigrow startup command to the proserve of the database to avert the performance penalty of growing the bi during crash recovery.
The database will then go through crash recovery on restarting. Thereafter, schedule an investigation of the hardware and OS for faults.