Moving a PVC/Suite instance to a lower XTM version


Sometimes you might want to revert your XTM TEST instance to its lower/previous XTM version, for one reason or another. For example, the current version of your XTM TEST instance is 13.4, but a recent bug fix that was specially targeted to version 13.2 has just been deployed on that version. Your XTM PROD instance is on version 13.2, but you first would like to perform tests of the said bug fix on the dedicated TEST instance.

The process of “downgrading” an XTM instance is a complex task, consuming a substantial deal of time and resources. The next section will list all the implications related to this operation.


Keep in mind that XTM International strongly advises that you do not revert an XTM instance to a lower XTM version, irrespective of the cause. You should only consider doing so as a last resort!

Furthermore, the process can be performed only on TEST XTM instances! Reverting PROD environments is out of the question.

Naturally, all Cloud-based instances (situated on XTM Cloud) are also excluded.

What are the implications?

First things first, keep in mind that XTM International does not offer any kind of feature for “downgrading” a particular XTM instance. The operation itself is a manual task that consists in reinstalling an XTM application on the server from scratch. For this reason, there are a number of factors/implications that you need to be aware of and accept. There are as follows:

  1. The duration of the whole operation depends on a server hosting provider, at which time the server is not accessible to anyone:

  • OVH (OVH Groupe SA): 8 hours,

  • AWS (Amazon Web Services, Inc.): 4 hours.

  1. During the process, all the data that exists on that XTM instance is irretrievably lost, such as projects, users, translation memory (TM), terminology, filter templates, SSO configuration (if it exists), and database.

  2. Since all the data is erased, you might probably want to have the database from your XTM PROD instance copied to the TEST instance. Depending on the overall size of the production database, this requires additional time – from 1 to 4 hours.