Checking feeders

[nhavis]

I check a rule calculated cell.
It is being calculated properly and seems to be fed fine, it shows up on zero-suppressed views, consolidates up etc.

When I edit or remove the feeder for that rule calculated cell, I check to see if the cell is still fed (which it shouldn't be), but there are no signs of such.
It still shows up in zero-suppressed views, it still consolidated up, and check feeders does not say 'unfed'.

Is the calculation flag still on this cell? Why?

A logout and login did not stop the cell being fed, however a server restart did.

What's going on here?
How can I change a feeder and validate whether or not it still works?

TM1 9.4, Windows XP SP 2.

[Alan Kirk]

I check a rule calculated cell.
It is being calculated properly and seems to be fed fine, it shows up on zero-suppressed views, consolidates up etc.

When I edit or remove the feeder for that rule calculated cell, I check to see if the cell is still fed (which it shouldn't be), but there are no signs of such.
It still shows up in zero-suppressed views, it still consolidated up, and check feeders does not say 'unfed'.

Is the calculation flag still on this cell? Why?

A logout and login did not stop the cell being fed, however a server restart did.

What's going on here?
How can I change a feeder and validate whether or not it still works?

TM1 9.4, Windows XP SP 2.

I haven't tested this myself, but I'd regard this as being expected behaviour.

For TM1 to go through and "unflag" every cell that is no longer fed when you remove a feeder... well frankly I'm not sure that it's even possible since it would need to keep a snapshot of every cell that is fed by the feeder, and then compare it to what the current set of feeders are doing after you recompile the rule.

So how long would you like your server to hang for while it does that?

(Either that or dump and reload the feeders as if it were doing a server restart... which is a less than pleasant thought for those of us who have servers which take over 15 minutes to come up because of heavy use of feeders.)

It may be annoying if you want to do diagnostic checks on your feeders, but for operational purposes there's no real harm in having the feeder flags remain in place after the feeders themselves have gone and there are certainly performance advantages in TM1 not having to go through and unflag everything.

A server restart is another matter, of course, as you've discovered. Only the cells affected by the new feeders will be "flagged".

[Gregor Koch]

Hi,

If you right click on the cube in question and choose 'unload' and then double click on it you can save yourself the server restart. 'Unload' will drop the cube (including the feeders) from memory. The double click will reload it into memory and shouldn't take much longer (unless you have gigabytes of stored values) than a rule save on that cube.

Trickier with intercube feeders but also works.

Cheers