Loading an existing Cube into TM1

[Riposte]

Hello,

I copied an cube (folder\files) from our old TM1 Server to our new one. Now I would import the cube into the TM1 server on the new server.

How those this work? Is there an import function?

Thanks.


Best,

Riposte

[Alan Kirk]

Hello,

I copied an cube (folder\files) from our old TM1 Server to our new one. Now I would import the cube into the TM1 server on the new server.

How those this work? Is there an import function?

If you've put the cube's .cub file (and any rules files, and any dimension files, and any dimension attribute files) into the new server's data directory, all you need to is reboot the new server and it will load the new cube up with the original cubes.

[David Usherwood]

In my experience you just need the .cub file and the .dim files (some of which you may already have). If you use an .dim file which is already in the server folder and it is missing elements which the .cub file expects the data will go but the remaining content will appear.
But I agree, to make it work you need the rest of the content and you will very likely need to adjust for inter-cube rules etc.

[TrevorGoss]

Why does no one ever recommended Performance Modeler for this sort of task? The transfer in and out process is very useful for this sort of task? Is Performance Modeler that bad? Is it so unloved by TM1 experts that it never deserves a mention?

[Riposte]

worked - thx.

Take care that you have set the correct directory path in the Tm1s.cfg file für data and log directory if you move cubes between servers

[Alan Kirk]

Why does no one ever recommended Performance Modeler for this sort of task? The transfer in and out process is very useful for this sort of task? Is Performance Modeler that bad? Is it so unloved by TM1 experts that it never deserves a mention?

To your last two questions, yes and yes. If it was a wizard instead of software, it's name would be Voldemort.

But in all seriousness, the reason that I didn't mention it in this particular case is that the original post implied that the files had already been moved to the new server's data directory, so going through the process of creating a transfer in Muddler seemed to be a lot more time consuming than just rebooting the server to pick them up.

[lotsaram]

Is Performance Modeler that bad?

Yes.

Is it so unloved by TM1 experts that it never deserves a mention?

Maybe it isn't so bad not to mention it.

To elaborate on why Performance Modler is so bad let me just mention what makes it unusable from my POV.

On launching the Modler client it sucks every last piece of metadata down from the server to the client. That's the full text of every single TI process, full text of all rules, every view definition, every member of every subset. You see a problem here? Why not just retrieve the list of objects and the detailed contents when needed by the user? ... well that would be just too sensible. The end result is that although Modler may be just fine to fire up on PlanSamp with all of 4 cubes and 2 TI processes point it at a decent model with a few hundred objects and you can not just grab a coffee while you wait you might as well go pick, roast and grind the beans as well. This renders the product more or less useless as developers need to be productive not spend their day waiting for editors to load.

I really like some aspects of it, I would like to use it more if the flaws got fixed, but the flaws are just too many.