Parallel Load for users without scidb admin account


I wonder if you have any plan for changing the restriction that parallel loading must be done by scidb user? It is a bit inconvenient because not many users can have scidb admin access, and I don’t see the reason why this must be a restriction, either because of technical difficulty or uncommon usage from non-scidb users (this is big data analysis and everyone has to move big data). Currently we have custom solution, but it is always nicer to have support directly from the system itself.



It is not strictly true that the loading needs to be done by the scidb user, although it is more convenient. The actual requirement is that the user must be able to write to the storage directories of each instance, on all nodes. We do this by ensuring that each authorized user is in the scidb group, and changing the permissions on all of the directories after initialization to ensure the group has write access. A bit of an administrative burden, but that is what sys-admins are for. This also is not the best solution security model-wise, but that always comes down to trusting your users anyways.