Replication database in Tamino 4.1


I have read the documentation about “Replication Guide” of Tamino 4.1 and have got some question about it.

Now I use old version 3.1.1 (and 3.1.2)

The most important things personally for me is if when this replication will start automatically to work when the primary system fails? I just want to have to make replication process start always when is updating, modification, adding, deleting processes. Is it possible in Tamino 4?

For example I have got 2 db tamino on Windows 2000 Pro in first productive environment and I want to have 2 db replication tamino in the second environment. I want that when the primary system of database fails (productive environment), the additional database (on second environment) will be then databases which makes application to work. Is it possible to make this process in automatic way?

The second questions is about system. Does the replication work on:
1) MAINFRAME system ?
2) SUN SOLARIS system ?
3) UNIX system ?
4) LINUX system ?

Is there any difference about implementation on these four different systems? What about comparision to Windows 2000 Professional?

These all above information are very important.

Any help will be appreciated.

In advance, thanks a lot!

best regards,
Dariusz Baumann

Hello Dariusz,
here are the answers to your questions (as far as I understood them):

0. “Replication” is not to be mixed up with “High Availability”.
Master and replication database must both be active in order to “pull over” the updates from the master and reapply them.

1. Replication works “by name”, names must be and stay unique.
When the master database (e.g. “master_db”) fails and the replication database (e.g. “repl_db”) shall take over its role, the following steps have to be performed:
The (former) master must be deleted or renamed (which however is only possible if it’s list of replication databases is cleared)
i.e. “Delete Database” : master_db
or “Start Database” : master_db
“Remove” repl_db from master_db’s permission list
“Rename Database” to a name which is not know to the applications: e.g “master_db” -> “former_master_db”

The (former) replication database must be switched to become a normal database and renamed to the (former) master’s name
“Stop Database” : repl_db
“Set Type to Normal Database” : repl_db
“Rename Database” to the former master’s name (otherwise the applications will not find it): “repl_db” -> “master_db”
“Start Database”: master_db (the former replication database now has the former master’s name!)

If you now want to set up replication again, you have to start from a non-initial backup of this new master:
“Create Backup” of this (new) master

“Add Replication Database” to this (new) master

“Create Database from Backup”: a (new) replication database from the non-initial backup of the (new) master_db
“Set Type to Replication”: (new) repl_db
“Start Database” : (new) repl_db

Applications can now run unmodified against the (new) master database.

2. Replication is available on all platforms where Tamino 4.1.1 is released

Hope this is helping you.


  1. Replication works “by name”, names must be and stay unique.
    First all of, thanks Uli Faltermann for very quick reply.

    However, I wonder if there are the master db and the replication db on different machines, the names of databases do not have to stay unique? It it right?

    best regards,
    Dariusz Baumann

Hi Dariusz,

You are correct, the databases names can be different. The Tamino versions for the master db/replication db must be v4 and must reside on the same type of operating system as well. This is because the backups files (which are used for the replication) are binary which causes problems for different architectures.


Stuart Fyffe-Collins
Software AG (UK) Ltd.

Hi all, I try to replicate a database that is in different nodes (machine), but I don’t know how indicate that the master database locates in different node. Do you tell me how must I do?
Thanks in advanced.


Hi Esteban,

sorry for the late reply. Tamino finds the master database by name, regardless of the node it is running on. The prerequisite for this is that you use the same XTS directory service for master and replication database.