I have to apply fixpack on DB2 V9.1 FP5 to upgrade it to fp7. The database has HADR implementation.

I know about the rolling fixpack option but the team wants to shut down the HADR completely on both primary & standby, apply the fixpack and bring up the hadr.

The steps i can think of are: -

Stop HADR ON peer:-
db2 deactivate db zinga
db2 STOP HADR ON db zinga
db2stop

Stop HADR ON primary:-
db2 STOP HADR ON DB zinga
db2stop

Apply fixpack on both the servers/update instance/das update

Start HADR ON peer:-
db2 START HADR ON DB zinga AS STANDBY

Start HADR ON primary:-
db2 START HADR ON DB zinga AS PRIMARY

Let me know in case I am doing something wrong since I am not sure with the steps right now.

You can't stop HADR completely without using a backup and restore to re-establish the standby. I am not 100% sure about 9.1, but you should be able to:

1. deactivate db on standby
2. db2stop on standby
3. db2iupdt on standby
4. db2start on standby
5. activate db on standby (this should casue HADR to be Peer mode again)
6. Takeover HADR on standby to switch roles
7. deactivate db on old primary (now standby)
8. db2stop on old primary (now standby)
9. db2iupdt on old primary (now standby)
10. db2start on old primary (now standby)
11. activate db on old primary (now standby)
12. Takeover HADR on standby (old primary) to switch roles back to original state

Even if the application is stopped for you to do an upgrade, I don't see a problem with Marcus_A's approach. There is no need to shut down the servers.

When shutting down the databases/instances, you should always do the primary first. This will ensure that no transaction happens after standby is shut down.

HTH

Sathyaram

--

"You can't stop HADR completely without using a backup and restore to re-establish the standby. "
There is no problem in stopping and starting hadr completely. as long as you do not bring the standby out of rollforward - no problem. we even have tsa and on accept we stop the both servers and apply fixpack and restart all - without backup/restore..

The first procedure requested can be followed without problem..

Hi There,
In the thread listed below: -
Apply fixpack on HADR Servers
you wrote that you did stop the HADR on both servers,apply fixpack and restart all.
Can you please list down the steps to be taken on primary/standby server for it. It is quite imp for me.Thanks in advance
Yes, my salary is also very important to me. I did stop spoonfeeding long time ago. All info is in the book, so at least : have a look and try to understand, OR hire somebody that does and could inform you how to do this in the future..
Maintaining a cluster domain using db2haicu

I agree 100% with you, Guy.

Personally, I don't mind the very basic questions even, as long as it shows that they have put in efforts to understand, like this one

Why Am I an Error with this Query?

But, the ones below are annoying.

Join Query - Urgent!!!
Query is Taking too much of time around 30 mins... plz help me out folks :-(

-