I think the DB is beyond repair. To work within a space constraint on this
play db, I have a cron job to purge any LOG that's older than 7 days. I'm
guessing the LOG needed to complete the new container request was already
deleted by the cron.

To change topic slightly,

1. is there a way I can make the cron job more intelligent so that it will
purge only non active logs?
2. is it true that DB2 will continue to "function" even if the active log
was deleted?
3. how true is the beyond repair assumption?

Thanks for your ideas in advance.

Cheers
Ray



Raymond Lam/RBA
04/25/2007 03:49 PM

To
"LazyDBA Discussion" <[Email address protected]
cc

Subject
SQL1036C An I/O error occurred while accessing the database.





hi gurus - thank god I'm getting this error on a play database but it'll
be great if I can learn the steps to fix besides restoring from backup.

What I'm trying to do is to compare the I/O performance between a
container located in NETAPP Storage versus LOCAL hard disk.

So I started with creating a new tablespace via a 3rd party GUI tool. For
some reason, the generated script was hanging and so is "LIST TABLESPACES"
The file SQLTAG.NAM with 0 byte was created and I had to kill the instance
to terminate the application holding the lock.

Now that the instance was restarted, I could no longer connect to the
database due to the I/O error. Is there a way I can fix up this play db?

CREATE TABLESPACE LOCAL4K
IN DATABASE PARTITION GROUP IBMDEFAULTGROUP
PAGESIZE 4K
MANAGED BY SYSTEM
USING
('/usr/tmp '
)
EXTENTSIZE 32
PREFETCHSIZE 32
BUFFERPOOL DATABP
OVERHEAD 24.10
TRANSFERRATE 0.90
DROPPED TABLE RECOVERY OFF;

Cheers
Ray

DB2 & UDB email list listserv db2-l LazyDBA home page