Hello friends!

I have a question:

I installed two servers with DB2 V9.5 and HADR Functionality. Automatic Client Routing and Takeover function works fine. But there havent been any loads on my primary database.

To avoid an occuring problems, i want to know, what can i do to make my database highly available, even if there will be some loads on my primary.

The IBM Information Center is not totally clear about it. Can anyone explain me this?

Thanks in advance.

we have been using hadr and tsa intensively on our systems
we use load - oltp ..
there have only been issues with tsa during test-failover and test-disasters
what do you mean exactly - can you be more precise ?

My HADR runs with HACMP.

I just want to know what can I do, in case there will be a load on my primary database, to avoid a problem.
Right now I understood the Load HADR Problem like this:

If there is a load on my Primary database, DB2 is not able to replicate the data to the standby.

That would mean that, after a load happened on the primary, I have to make an Offline Backup ( which means downtime ) of my primary, restore the standby database and run HADR Process again...

Can I avoid this downtime somehow?

you have to use load with copy yes option
put the copy file on a shared filesystem between prim and standby or use tsm if available

that means if i have two aix systems, i mount from standby one folder on primary where the copy file is in.
and then i run " db2 load from file.ixf of of ixf copy yes" on primary

Right?

yes, as long as copy file can be created by primary and used by stdby for load from same path = ok
see for details :
Database configuration for high availability disaster recovery (HADR)

The easiest way is use IMPORT instead of LOAD. IMPORT uses INSERT logic, so they will be logged and show up on the HADR standby. IMPORT will take longer than LOAD, but unless you have hundreds of million rows it is usually not too bad. Make sure you set COMMITCOUNT to about 5000 on the IMPORT command or else you will run out of log space.

Okay that helped me ...

I will see then if its possible to use the Import Command instead of the Load Utilitiy :-)

Thank you guys.

Just one more question: if Import is not possible, what is the easiest way do make my HADR ready for any loads?
Does anyone have a "best practise"?

as indicated before - change db cfg setting to log index build and use copy yes option with load.. the infocenter explains this in detail.. (see for hadr)

Heureka, I finally found what I was looking for. Since Version DB2 Version 9.5 Fixpack 4 there is a new DB CFG Parameter "BLOCKNONLOGGED".

For more informations, check this out:

IBM DB2 9.5 Information Center for Linux, UNIX, and Windows