TM1 Forum

Guys,

I have built some security rules. The security needs to change as a status changes (controlled in a lookup cube). The rules work but only if you refresh security. Is there a way (apart from using TI) of setting security to refresh autonatically??

Thanks,

Jim.

jim wood wrote:Guys,

I ahve built some security rules. The security needs to change as a status changes. The rules work but only refreshing security. Is there a way (apart from using TI) of setting security refresh to auto??

Thanks,

Jim.

If you have security rules then you need a security refresh.

If security cube values are populated via TI (via ElementSecurityPut for example) then no security refresh is required, ... it can save a lot of time. For large models it is worth considerign converting to TI maintained vs rule maintained element security.

I'm building a real time work flow system managed by security. This is why I said excluding TI. What we are thinkning of doing is building in to the submission procees (done via the web) a batch that executes a refresh TI process,

Jim.

I have amended my original post to help avoid any future mention of TI!!!

jim wood wrote: I have built some security rules. The security needs to change as a status changes (controlled in a lookup cube). The rules work but only if you refresh security. Is there a way (apart from using TI) of setting security to refresh autonatically??

No.

Since there is no (documented) server config parameter for this, your options are:
(a) One that you don't want to hear about but which would probably be the most desirable option for the reason that Lotsaram mentioned given that security refreshes can slow the crud out of a system. Multiply that slowness by any number you like if you have them as part of a workflow. (And I'd guess that that's why there is no such documented parameter);
(b) Build a call to the relevant API function into the batch file that you refer to; or
(c) Train a chimp to bang on the GUI menu item at regular intervals.

This is another instance where the solution that you don't want to hear would be twice as useful if they could be fired by event triggers, but you go to war with the OLAP tool you've got, not the one you'd like to have.

jim wood wrote:I'm building a real time work flow system managed by security. This is why I said excluding TI. What we are thinkning of doing is building in to the submission procees (done via the web) a batch that executes a refresh TI process,

Jim.

Well you said you didn't want to hear it ... but here it comes.

If your real time workflow can include a "submit" button then you can easily run a parametized TI to read from your control cube and populate the security cube(s). It will be lightening fast as in the user won't realize anything is running in the background, and it would likely give a much more "real time" result then needing to run a security refresh.

Hi,

Did you try writing rules on the control security cube ?

Like if you want to implement element level security. Write rules on }Element_Security_**** cube.

These rules wont require refresh.



Thanks and Regards,
Sushanti

I would avoid cell_security for very large cubes and use security cube based rules- this is because in my experience the mem usage can balloon if you have a large amount of cells in your cube

and I know that you don't want to hear it but if you are already kind of using TI why not continue to use it? easier than trying to make API calls

I'd go with lotsarams second suggestion, its what I've done in the past and has worked better than cell based security (which i tend to use for smaller cubes)

i've also created my own workflow in TM1 before and used TI, Security Rules and Lookup cubes so you are on the right track

sushanti wrote:Hi,

Did you try writing rules on the control security cube ?

Like if you want to implement element level security. Write rules on }Element_Security_**** cube.

These rules wont require refresh.



Thanks and Regards,
Sushanti

Hi Sushanti,

I think you can safely assume that Jim knows how to write a rule in a security cube. The whole point is that when using rule based security a security refresh is required for changes to take effect within the security model. Maybe you haven't discovered this yet but it is indeed true.

Oh ... I didnt face this refresh issue though the cube is large in size....

Can u tell me 1 thing, R u using SaveDataAll TI Process to save data frequently ?

Thanks guys. As I said above adding the TI process to the submit process is the only way I could think of doing it. The API might be another option as we will be already going through it for some of the .net functions that we are building in.

Thanks for all your posts,

Jim.

sushanti wrote:Oh ... I didnt face this refresh issue though the cube is large in size....

Can u tell me 1 thing, R u using SaveDataAll TI Process to save data frequently ?

The model we are building will be used by the client across the world. For this reason we are keeping saves to a minimum to reduce down time. We may do it once a day.