Hi all,
Is there a way to set up HADR when primary db is in version 8.2
and the standby db is version 9.5 ?

When i take an ofline backup from the primary db and restore it to the standby db ,it is not in rollforward pending state and i get error message:

$ db2 RESTORE DATABASE xxx FROM . REPLACE HISTORY FILE WITHOUT PROMPTING
SQL2555I The restored database was successfully migrated to the current
release.

db2 UPDATE DB CFG FOR ......

db2 START HADR ON DATABASE xxx AS STANDBY
SQL1767N Start HADR cannot complete. Reason code = "1".

Thanks !!

rc=1 means db is not in rollforward pending mode
try connect : you should see the message if in rf mode
otherwise regular prompt
see message :
1 The database was not in roll forward-pending or roll forward-in-progress state when the START HADR AS STANDBY command was issued.

doc indicates :
The primary and standby databases must have the same operating system version and the same version of the DB2 database system, except for a short time during a rolling upgrade.

Ok Guy,
Tks,

They can be different DB2 fixpacks, but not different DB2 versions, since the transaction log formats are usually not indentical between different DB2 releases.

Just a small remark Marcus. Logs are not being shipped - only logbuffer..
and how about rolling upgrade .. you can not break a hadr when active ..
it is not permanent, only untill both servers have been upgraded ..

I am not sure what is the distinction between log buffer and logs on disk. The format of the logs usually changes from one release to another.

I don't understand your other remarks. One can upgrade the standby to a later DB2 fixpack in the same release and then HADR takeover to the standby, and then repeat on primary server. I have done this many times.

Without getting into all the details, the following basic steps can be used:

• explicitly deactivate standby database
• stop instance
• upgrade instance
• start instance
• activate database
• if database does not automatically go into HADR peer state when activated, then do a start hadr on db as standby, and it should go back into peer state when catch-up is done