Steps to reproduce

1. Perform a full site export of a Confluence instance
2. Set up an instance with a different home directory (this is so that the backup path will differ from the export)
3. Check the backup path of your instance on the Backup Administration screen (/admin/dailybackupadmin.action)
4. Now import the backup you created in step 1
5. Check the backup path of your instance on the Backup Administration screen - it still shows what it did in step 3
6. Perform another site backup, with the option to archive a copy of the backup checked
7. You'll see that the temporary backup is created where you would expect (in the location that matches the backup path you saw in steps 3 and 5), but the archived backup is created in the home directory that was specified in the instance you imported from

In cases where that location doesn't exist anymore, a stacktrace will appear instead:

java.io.IOException: No such file or directory
	at java.io.UnixFileSystem.createFileExclusively(Native Method)
	at java.io.File.createNewFile(File.java:883)
	at com.atlassian.core.util.FileUtils.copyFile(FileUtils.java:465)
	at com.atlassian.core.util.FileUtils.copyFile(FileUtils.java:401)

See attached files for full stack trace.

I think the desired behaviour in this case is that the existing backup path of the instance is used, instead of the path from the import. This will ensure that the location always exists (since the home directory of this instance has to!).

Workaround

On the Backup Administration screen, click to edit the backup settings. You won't be able to change the backup path, but once you submit the form Confluence will go back to using the backup path that's specified on this screen.