Can I load a .cub file into TM1 without restarting?

[davidh]

Hi all.

Quick question.

Is it possible to load a backed up .cub file (with a new name) into a running TM1. Or is the only way to shutdown and restart TM1?

Thanks is advance


David

[Alan Kirk]

Quick question.

Is it possible to load a backed up .cub file (with a new name) into a running TM1. Or is the only way to shutdown and restart TM1?

For all intents and purposes you would need to reboot.

The server normally loads all of the cubes into memory at startup, though particular cubes can be flagged as "Load On Demand". However the server still knows about those; it loads them into the cubes list at startup even though it doesn't load the data. That's quite different to suddenly finding a new cube on disk and loading it, which it won't do.

There are also some third party API-based products which can "hot promote" a cube from a development environment, but in reality that means creating the cube on the server and copying the data into it. Again, this isn't the same as simply loading a disk-based .cub file.

[davidh]

Ok thanks, that's what I thought, was just holding out for hope otherwise.

[Alan Kirk]

Ok thanks, that's what I thought, was just holding out for hope otherwise.

Of course the other kludge would be to load it into another environment and to use replication to suck it into the main server, but I'll leave Matt Goff to wake screaming "NOOOOOOOOO!!!" from his nightmare about that one.

Frankly a reboot would be easier and cleaner.

[Steve Rowe]

Actually.....this can be done if you have a different version of the cube already running in the TM1 Instance.

Unload the cube from server explorer (right click on the cube).
Replace the .cub on the disc with the replacement.
Load the cube by opening a view, this may take time as it will trigger the feeders for the cube and note that if you have feeders into your target cube from other cubes these won't have executed and the cube incorrectly fed.

You should get the revised data, though if the dimensions have changed a lot you will have stuff missing.

Obviously a bit of a hack but might get you out of a hole at some point...

Cheers,

[Alan Kirk]

Actually.....this can be done if you have a different version of the cube already running in the TM1 Instance.

A nice piece of lateral thinking, there!

[macsir]

Actually.....this can be done if you have a different version of the cube already running in the TM1 Instance.

Unload the cube from server explorer (right click on the cube).
Replace the .cub on the disc with the replacement.
Load the cube by opening a view, this may take time as it will trigger the feeders for the cube and note that if you have feeders into your target cube from other cubes these won't have executed and the cube incorrectly fed.

You should get the revised data, though if the dimensions have changed a lot you will have stuff missing.

Obviously a bit of a hack but might get you out of a hole at some point...

Cheers,

Hey, Steve. Is it still true in PA? I know this was long time ago. I tested it in now and looks like it is not working. No data is showing even if the backend cube file has been replaced. Thanks.

[Steve Rowe]

I've not attempted for a long while so it may no longer be possible.

If the cube is fed by another cube you'll need to process the feeders for the feeder source cubes as well.