Add additional step in Content Anonymiser documentation

Re: https://confluence.atlassian.com/display/DOC/Content+Anonymiser+for+Data+Backups

Following this process to clean entities.xml, rename it to cleaned.xml and re-archive it into a backup file results in an unusable backup file because the restoration process expects a the xml file to be named entities.xml. I tested this in versions 5.8.10 and 5.9.1.

IF the purpose of this document is to create a new backup file with anonymised content (which is not clear in the documentation), then after step 4 we should add an additional step, something like:

4.5. Backup the previous version of entities.xml to a new location, and rename cleaned.xml to entities.xml

And then step 5 should be modified to reflect that new step:

5. Re-ZIP the new entities.xml with its exportDescriptor.properties (and any other backup files from the archive) to ensure we (eg. Atlassian Support) know exactly which version of Confluence the XML backup came from.