When downloading files via either the SpiderOak One web interface or the Management Console of SpiderOak Groups or Enterprise, you may get what identifies itself as a server error. That usually means that one of the files in the download set had been backed up in a corrupt state.
To identify which file is causing the download to fail, winnow down the download set with successively smaller selections such that eventually you will successfully download everything except the corrupted file (which being corrupt will not itself be downloadable). As an example, say you want to download all of:
A A1 goodfile A2 goodfile B B1 goodfile B2 BADFILE C C1 goodfile C2 goodfile D D1 goodfile D2 goodfile
On your first try you attempt to download A, B, C, and D all in one go. That will fail due to the corrupt BADFILE, but you don't know which file is corrupt. So on your second try, split it into two downloads: A+B (which will fail) and C+D (which will work). Then split the failed download into two parts, A (which will work) and B (which will fail). And so on. This will take multiple passes to complete (particularly if there is more than one corrupt file) but it gets the job done.
As you do so, keep in mind our recommendation to use download selections of 10 GB or less at a time.
Moving forward, prevent this problem by not directly backing up things known to get corrupted when directly backed up. Examples include live databases, software development environments, and application data. For our full recommendations on what to back up with SpiderOak One and Groups and what not to, see What Data to Select for Backup. Also do not select an entire volume root to be backed up.