since some time a fast copy process is active to copy data from the beamline file system to the core file system immediately after data are written into the beamline file system. As there appeared some missunderstandings due to poor communications from my side some clarifications are needed: 1) The new fast copy process is only active for BEAM TIMES, *NOT* for COMMISIONINGs. 2) The old periodic copy process is still active for everything, i.e. beam times and commisionings. For beam times it will pick up any files that were not copied by the fast copy process, and copy tham as it has always done. 3) The fast copy process mirrors renames and deletions in the beamline filesystem to the core file system, which is intended to keep properly handle cases where data are written to e.g. a file ending in *.tmp an when finished renamed to it's final name, or to cope with the way editors handle backup files. 4) In case of hickups on the proxy nodes, e.g. when a node dies or has some other problems the fast copy process running on that node will also fail, there is as of now no high availability mechanism in place to start the process automatically on the replacement node that takes over the NFS/SMB/Hidra funtionality. In that case the old copy process is the fallback! It is not considered good practice to write data into files that always have the same name and rename them after writing to some meaningful unique name, or to write always into a folder with the same name and then rename the folder to a meaningful unique name. While this may seem to be a good idea as one does not have to change file names in the data taking software, it's really a recipe for disaster. If data taking software insists on fixed file names or is unable to use unique ones like containing the time as part of the name then that software is badly broken and needs to be fixed. So it is STRONGLY DEPRECATED to rename files and folders in the beamline file system, rather choose proper names when creating files or folders!