we are using db2 v8.2.3 aix 5.2

is it possible I setup HADR in one serever between two databases that are located in different instances?

thank you for yourhelp.

I don't see why not

Thank you n_i for your answer,

I have read some article about HADR and all talking about primary and stand by servers, I know replication is possible but I did not know even HADR is possible.

how can I describe failover on stand by server? or rerout client to standby server? this practice is just for myself and I want to learn more about HADR.

http://www.redbooks.ibm.com/redpiece...7363.html?Open

is it true in HADR we should have same name for instances and log files in primary and stand by servers?

Instance names don't have to be the same.

If you run both primary and standby on the same machine then obviously you cannot specify the same location for log files; they will have to be relative to the each database directory.

By the way, don't you think that running the primary and standby databases on the same server kind of defeats the purpose of HADR?

you are absolutly right, unfortunatly there are not two servers available in our environment and I want to setup that just for myself and I get some knowledge about this feature.

Initially, when Tivoli System Automation was released and included for free with HADR to automate HADR failovers, the instance names had to be separate in order for the HA scripts to work. But that is not longer true with TSA 2.1+.

Regardless of whether the instance names are the same, all the tablespace container paths have to be the same on both servers (so that create/alter tablespace SQL will replicate properly), which is easier if the instances are the same.

Also, there is an unusual situation where a C user defined function is executed in a stored procedure, and the load module of the function may not be found on the standby if the instance names are different (I ran into this problem). So making the instance names the same is definitely better.