TM1 Forum

Dear all,

I have one question:

There is a TM1 system which has 128 cubes built in, when I try to log in, it takes more than 2 minutes to load those cubes, I just wondering that is it any method to force those cubes do not automatically loading, only when I click the cube it starts loading instead.

Thanks for your help!

kiss351523 wrote:Dear all,

I have one question:

There is a TM1 system which has 128 cubes built in, when I try to log in, it takes more than 2 minutes to load those cubes, I just wondering that is it any method to force those cubes do not automatically loading, only when I click the cube it starts loading instead.

Thanks for your help!

I have no idea about this, but I also want to know this answer.
I think you can modify the subject to such as" how to force the cubes don't run automatically", thing like that.

I'm no google master, but "tm1 load on demand" gets me this. It's a clue.

http://publib.boulder.ibm.com/infocente ... 81325.html

If we talk about the time it takes to just log in to TM1 using a pure TM1 interface, then Load on demand has no effect on it. This parameter simply moves the cube load time from server startup phase to the point when you try to open the cube (the click/double click on the cube ).

Furthermore, if I assume your TM1 server is already up and running during the test, then you should get the list of objects (incl. cubes) displayed immediately after login.

The critical issue here is that a slow network will cause such a delay, which is valid for clients like Architect & Perspectives, while TM1 Web can still work properly.
So, the primary question is, what type of client are you talking about? Having the chance to connect remote to the server box and start Architect there would give the quickest answer to the network topic.

tks for your help!!

Gabor wrote:If we talk about the time it takes to just log in to TM1 using a pure TM1 interface, then Load on demand has no effect on it. This parameter simply moves the cube load time from server startup phase to the point when you try to open the cube (the click/double click on the cube ).

Furthermore, if I assume your TM1 server is already up and running during the test, then you should get the list of objects (incl. cubes) displayed immediately after login.

The critical issue here is that a slow network will cause such a delay, which is valid for clients like Architect & Perspectives, while TM1 Web can still work properly.
So, the primary question is, what type of client are you talking about? Having the chance to connect remote to the server box and start Architect there would give the quickest answer to the network topic.

TKS for you help!

Hi
I fully agree with Gabor. We had a system with a very fast WAN connection where all query times, drill downs in cube views, DBS formulas etc. worked within seconds, but during login, simply listing all server explorer objects (cubes, dims, processes, etc.) took for 5 minutes. We couldn't do too much about it unfortunately (other than workarounds like terminal clients, web, etc.)

So it can happen that even with ana cceptable WAN speed the listing of server objects is a way too slow.
I also had another thread about this issue here on the forum

One question...do you have 'Display Control Objects' on? If you do, try turning this off to see if it may help.

Another interesting artifact that I have noticed that leads me to believe the issue is actually the Perspectives/Architect UI not TM1; but I have not really investigated it deeply. If you perform a login with control objects on, and say login takes 10 seconds. To turn OFF 'Display Control Objects' now takes 40-50 seconds. This indicates that the problem may not be TM1 necessarily, but how the UI is handling adding and removing nodes from the tree control they are using.

Again, I have not investigated this deeply because my work around is log out, then turn off 'Display Control Objects' and then log back in. That seems to be much faster, though more clicks.