Issue Summary

After setting up a Disaster Recovery as per the official guide, at some stage (see replication steps below), folders with stripped names will be created in the secondary location configured in JIRA

Steps to Reproduce

1. Spin up a 1-node DC instance.
2. Set up DR Replication, (including checking the checkboxes at least avatars & attachments)
3. Do Not click “Synchronize” (may not be important, but keeps things clearer for the test)
4. Shut down the node. double check that there’s not any “secondary” folder in the shared home (this is the default DR replication location)
5. Start the node up again and wait. Once the healthchecks fire, the “secondary” directory should be created with the spurious entries.

Expected Results

Only the folders from the shared home folder should be replicated to the secondary location, no spurious folders should be created in the secondary location.

Actual Results

Empty folders like below will get created outside of the 'data' folder:

Workaround

This is a minor bug which has no impact on the actual replication, hence can be safely ignored