SecurityAdmin - bending the rules
Posted: Thu Dec 15, 2016 2:33 pm
I have a client with a SecurityAdmin that needs to administer 9 TM1 instances across 3 servers.
Users login via intregrated login and come from three different domains.
To facilitate the multiple domains and username variants, we have attributes on the }Clients dimension.
I then use a rule to test if the SignOn and Domain attributes are non-blank then derive UniqueID based on the client/signon and domain.
This works well unless you are SecurityAdmin as you can add the user but cannot update the attributes due to the restrictions imposed on SecurityAdmin.
SecurityAdmin is able to run TI's either - whether or not Security Access is checked or not.
Creating and additional group with permissions to access client attributes or a custom cube which is then used to rule derive UniqueID also does not work as the privileges do not inherit/are limited.
I was considering a process that the user could execute outside of TM1 to do a TM1RunTi with credentials allowing update only of the relevant cube but this would obviously mean creating password files, TI's etc. etc.
I fully understand why security has been implemented like it has been but looking for some leeway to facilitate update of attributes.
Have I missed something?
Any ideas?
edit: forgot to mention that they use a domain ID e.g. WXYZ0123 as the Client name rather than a fully qualified name e.g. WXYZ0123@DOMAIN
Users login via intregrated login and come from three different domains.
To facilitate the multiple domains and username variants, we have attributes on the }Clients dimension.
I then use a rule to test if the SignOn and Domain attributes are non-blank then derive UniqueID based on the client/signon and domain.
This works well unless you are SecurityAdmin as you can add the user but cannot update the attributes due to the restrictions imposed on SecurityAdmin.
SecurityAdmin is able to run TI's either - whether or not Security Access is checked or not.
Creating and additional group with permissions to access client attributes or a custom cube which is then used to rule derive UniqueID also does not work as the privileges do not inherit/are limited.
I was considering a process that the user could execute outside of TM1 to do a TM1RunTi with credentials allowing update only of the relevant cube but this would obviously mean creating password files, TI's etc. etc.
I fully understand why security has been implemented like it has been but looking for some leeway to facilitate update of attributes.
Have I missed something?
Any ideas?
edit: forgot to mention that they use a domain ID e.g. WXYZ0123 as the Client name rather than a fully qualified name e.g. WXYZ0123@DOMAIN