Consultor Eletrônico

Status: Verified

GOAL:

Replication across platforms

GOAL:

Replication from UNIX to Windows

GOAL:

Replication from Windows to UNIX

GOAL:

Can you have production database on UNIX, standby on Windows?

GOAL:

Can you have production database on Windows, standby on UNIX?

GOAL:

Must the source and target database reside on the same platform?

FACT(s) (Environment):

UNIX
Windows
Fathom Replication
OpenEdge Replication

FIX:

The question of Replication across platforms is often asked, typically: a production database on UNIX with a standby on Windows (and vice versa), and how to replicate data between them. This Solution hopes to clarify the options available and our standpoint on this.

If the standby database is to be used as a hot spare (in the event of a crash, this is the database that is swapped over to), then the options are -

1. Application architecture design across platforms:
a. Trigger-based replication could be used, together with a means to enable and disable it (for example, so that the triggers no longer fire once the standby is switched to).
b. Alternatively, by implementing a means to tag updates to the production database (e.g. order numbers, order date/times), a 4GL procedure could be run as a batch job to check what the updates are, write these to a text file - then apply these to the target database. Although the following solution is in relation to DataServers, similar trigger logic can be used for Progress data sources: P63288, "How to Modify tables to create PROGRESS_RECID in support of the ROWID function in 4GL"

2. After-Image (AI) based replication
This is a very good "out the box" time-tested method. Since the introduction of OpenEdge 10.1A, the production database can be enabled for automated AI management. It should not however be used in a UNIX-Windows, Windows-UNIX cross-platform configuration. This method not only involves backing up and restoring across platforms but also copying the AI files then processing these files