Hello everyone,
We are currently migrating a GNUHealth/Tryton installation from GNUHealth 3.x (Tryton 5) to GNUHealth 5.x (Tryton 7).
The database contains:
-
~12k patients
-
~3k product templates
-
~4k products
We initially tried a direct Data migration using Proteus scripts, but we encountered multiple compatibility issues between GH3 and GH5, especially concerning:
-
product UOM mappings
-
category mappings
-
model evolution between versions
-
import stability for large datasets
We tried different strategy:
-
Export normalized data from GH3 into UTF-8 CSV files
-
Reimport them through Tryton 7 official CSV import mechanisms
But, we also have errors with some fields. The exported model is quite simple.
We would like feedback from the community on:
-
recommended practices for large GH3 → GH5 migrations
-
whether CSV import is preferred over Proteus for this use case
-
best strategy for preserving relations between templates/products/patients
-
known pitfalls concerning product.template and product.product migrations
Any advice or existing migration tooling would be greatly appreciated.
Thank you.