TM1 Forum

I have a cube in Instance 1 which should be dynamically associated to a cube which is residing in Instance 2. I went through lot of documents but was not able to find an approach to write a rule between two cubes residing on different instance.

Is there any approach through which we can write a rule associated between two cubes residing on different servers/ instances or TI process to copy data from one one cube to another residing on different server/instance.


Thanks i advance

You can't do that. The only way to link data between two instances of TM1 is through replication/synchronization or a TI process with flat file export import or ODBO, neither of which is dynamic.

Rule - no.

TI - yes. You will either need to use an ODBO connection (this is one of the available data source types in TI), or do a flat file push-pull approach. ODBO will let you connect directly, but you will find that a lot of people prefer the flat files method just because it's easier (no messy MDX), more reliable and, counterintuitively, tends to perform better. But I suggest you give ODBO a go, maybe it's a good approach in your case. You can also use Replication, but I would not recommend that approach - too unreliable.

qml wrote:too unreliable.

or to use my experience - shonky!

qml wrote:You can also use Replication, but I would not recommend that approach - too unreliable.

Also, the logs must be kept for the cubes in question, which is not always an option in my opinion.

Not wanting to be left out on this one, we have attempted to get rep and sync working on a few occasions and we got it working consistantly in one instance. It did however stop working as soon as we applied a fix pack and we couldn't get it working reliably again. Bascially as the others have said, avoid it.

Thanks all for you precious suggestion and being so kind. Looks like TI is a better option. Will try to implement it.


Thanks again !!

Steve Vincent wrote:

qml wrote:too unreliable.

or to use my experience - shonky!

Just to add to this one that my experience of ODBO in a production environment is that it is very shonky. It works OK for a small proof of concept but I've seen it crash the server (people logging in whilst ODBD process runs) too many times to trust it. This was back on 9.1.x or 9.4.x so the situation may have improved since then.

rmackenzie wrote:

Steve Vincent wrote:

qml wrote:too unreliable.

or to use my experience - shonky!

Just to add to this one that my experience of ODBO in a production environment is that it is very shonky. It works OK for a small proof of concept but I've seen it crash the server (people logging in whilst ODBD process runs) too many times to trust it. This was back on 9.1.x or 9.4.x so the situation may have improved since then.

Same experience here (on older versions, also never tested again on newer versions).
For replication between servers nowadays I would export data to a file and kick off a TM1RunTI to load the data on the other server. Seems the the most reliable option to me.

Michel