Unable to import Confluence Cloud backup into Server

[Simon Szumyłowicz]

Hello,

I've been trying to import a backup of Confluence Cloud taken yesterday (and one taken today) into a Server instance but it looks like there's some major incompatibilities between the Cloud and Server version. The error I got was:

Caused by: java.lang.ClassNotFoundException: com.atlassian.confluence.api.model.people.MigratedUser

Upon closer inspection of the `entities.xml` file in the backup and comparing it to a previous working one I've found some new objects not present in the old one which I believe are the cause for the import failure:

<object class="MigratedUser" package="com.atlassian.confluence.api.model.people">
<id name="key"><![CDATA[8a7f808969eb55e50169fe08f677077d]]></id>
<property name="username"><![CDATA[115e9f12d720d9675070caf292edf01f]]></property>
<property name="hashedUsername"><![CDATA[d3e5a4bb5cd842661999c9af085d7da1]]></property>
<property name="accountId"><![CDATA[12345:ab12345c-aaaa-1111-1a2b-12abcd12ab12]]></property>
</object> 

It looks like the Cloud version has some username obfuscation features which are incompatible with Server. Is my understanding correct that this will not work until the next Server release? If so - when is the release expected?

Thanks,

Simon

[Brant Schroeder]

Confluence Server and Coud have separate and distinct code base.  So a backup in cloud can not be recovered in Server.  You can migrate the data from cloud to server by following these steps.  https://confluence.atlassian.com/confcloud/migrate-from-confluence-cloud-to-server-724765578.html

[Simon Szumyłowicz]

Creating a cloud backup and restoring into server is exactly what the link you've provided (which I'm using as reference) says to do:

Export the data from your Confluence Cloud site, using the Confluence backup manager.

(...)

Import the data from your backup file (XML export) into your new Confluence installation (...)

Moreover this process works perfectly fine with an older backup (I believe it was taken a day before the 10/04/2019 feature release) which I have successfully used to complete an end-to-end dry run.