Has anyone ever experienced this?

We are getting this error and cannot connect to the database. Since DB2 cannot access the system/local db directory, a restore will not even run. We are trying to avoid a complete db rebuild.

The only similar occurence found was when running alternate fix-packs (which we are doing), was to disable the discovery for the 8.1 instances. The corrupted db is v 8.1.2 and not an 8.1 version, so this will be of no help.

Any ideas?

Thx

Try to uncatalog the server node: UNCATALOG NODE ...

Now,

Catalog the node again as a local node: CATALOG LOCAL NODE ...NODENAME..
(syntax check from command reference)
Now catalog the database at this Local node and try to reconnect.

Once I had the same issue on AIX (DB2 UDB ESE 8.1) and got that resolved like this way.

Thanks,
Jayanta Datta
Xansa India

Thanks Jayanta,

We wound up having to get DB2 support involved. Luckily, all that had to be done was to run a (db2demigdbd /home/db2home/sqllib 2) from a fp7 directory to move the db directory files back to a fp6 version. Once this completed, the db was once again connectable.

The key if you are running alternate fixpacks on the same server - do not login with a FP7 instance into a FP6 instance and run any commands. DB2 will convert the db directory files to FP7 ones and you will receive the SQL10004C error.

In the 'initial' fixpak 7 release, db2demigdbd has some issues and would not run properly ? Do you know if this has been fixed ?

Cheers
Sathyaram

Not certain, that issue was not discussed in the case.

But, here is more info on the fp7 version we are running :
Informational tokens are "DB2 v8.1.1.72", "s040914", "U498350", and FixPak "7"

Also, I had no problems when running the db2demigdbd, it worked the first time.

Thx