Recovering from SQL0293N without backups or logging

[sal...@il.ibm.com]

Hi
I am supporting a site where there was a problem where DB2 ran out of
memory. I attempted to fix it by increasing the utilities heap size. Once
I changed this parameter, I stopped and restarted DB2 and attempted to
reconnect. Reconnect fails with the message "SQL0293N Error accessing a
table space container. SQLSTATE=57048". I then checked the error log,
db2diag.log, and found that tablespace 8 is corrupted:

Error checking container 0 (/home/comet2/xlatA_2) for tbsp 8. Rc =
FFFFE60A

2001-01-31-13.47.50.898239 Instance:comet2 Node:000
PID:33544(db2agent (METEOR)) Appid:*LOCAL.comet2.010131112143
buffer_pool_services sqlbDMSStartPool Probe:800 Database:METEOR
DIA9999E An internal error occurred. Report the following error code :
"FFFFC11E".

Tablespace EV_XLATA_2 (8) is in state x0.
StartPool failed. rc=ffff c11e .

String Title:NOTE: PID:33544 Node:000
Leaving tablespaces offline at database startup requires recoverable
logging.

2001-01-31-13.47.51.215479 Instance:comet2 Node:000
PID:33544(db2agent (METEOR)) Appid:*LOCAL.comet2.010131112143
buffer_pool_services sqlbStartPools Probe:40 Database:METEOR
DIA9999E An internal error occurred. Report the following error code :

In fact, upon checking, there is no file /home/comet2/xlatA_2.
Unfortunately, no logging was enabled, no were there any backups of the db
made.
Log retain for recovery enabled (LOGRETAIN) = OFF
User exit for logging enabled (USEREXIT) = OFF

If I were able to connect, I would simple drop the tablespace, EV_XLATA_2.
I have spent most of today going through all the DB2 documentation, but
cannot figure out what to do. Without logging, I cannot use the RESTORE
command, and as I said, I cannot connect to drop the tablespace. I'm
stuck.

Thanks for your help.
- Eliot Salant

--
Posted from d12lmsgate-2.de.ibm.com [195.212.91.200]
via Mailgate.ORG Server - http://www.Mailgate.ORG