The backup process really has some unbelievable bugs.
If the backup fails, there is absolutely NO CLEANUP, leaving backup mount points and backup LV as is. Due to the incomplete backup, the next one will fail also (maybe partially, as remote backup rotation will not happen also).
This behaviour ate the complete storage of our backupserver, as now we did not have 3 backups in rotation, instead we had so many, that all storage was gone.
Saying that, we did the following:
1. removed all backup mount points and LV from the nodes.
2. deleted backups at the backup server
3. started backup run again in virtualizor.
same error, no change.
It connects to the backup server, it created the new directory, partially upload one or another vps and then it dies with the same ugly error message and no cleanup (or at least a warning message to an administrator) at all.