Hello,

My backup failed with sql0911n. I see this error:


2009-06-09-00.00.43.926292 I1598A523 LEVEL: Error
PID : 1960345 TID : 1 PROC : db2bm
INSTANCE: db2inst1 NODE : 000 DB : UATDB
APPHDL : 0-694 APPID: *N0.db2inst1.080470543
FUNCTION: DB2 UDB, database utilities, SqlubTableLock::request, probe:153
DATA #1 : String, 59 bytes
Error upgrading lock.
PoolId: 0, ObjectId: 2, Intent: 3.


Can you offer some advice? Thank you!

SYSIBM.SYSTABLES is locked.
Most common reason is RUNSTATS ?
Are you doing one, when you initiate the backup

Cheers
Sathyaram

Thank you. I don't think we execute runstats during the week. This is an online backup so why would db2 care if some table is locked? Any reason why it can't backup a table even if it's locked?

RUNSTATS takes a exclusive row locks in few system catalog tables ...

DB2 needs a consistent copy of the system catalog tables for backup. Hence, if the system table is locked (table or row) in exclusive mode, then backup fails.

If it is a user table that is locked, then backup will continue ...

Once the system tables are backed up, then it doesn't matter if RUNSTATS starts - as there is a consistent copy already for that image.

Hope this explains

depending what ver you are on, you might want to check on AUTO_RUNSTATS parm. If it is ON db is doing stats for you.

auto runstats is off.

Is there any workaround for this locking problem?

Yes, script your script to send a backup into a sleep mode for a while and then do another attempt.

Check for locks first and either wait or force them. there are more depending on your business requirements.

see the following technote for some additional info:
IBM - DB2 online backup fails with SQL0911N reason code 68

This is very common in our environment, most of time the deadlock was caused by runstats or reorg running on the table when backup was started. Also db2diag.log will show you the detail.
---------------
Eric Xu (许 昌)