TM1 Forum

Hi gurus

for one incoming project , I need to implemen parallel processing in the model , thus I am very keen to a explore when or what would cause the TM1 locking to happen . Would anyone share with your insight and reply to this thread ?

Let me be the first one .

1. Attribute Update would cause locking of all Cube objects ?
It seems to be the case in my testing , because whenever I try to update the attribute when a data load process is running , the attribute update thread would get a IXC lock . I guess that means it will get a gobal lock whenever attribute is updated.

Generally speaking, there haven't been any "global" locks in TM1 since the changes to the locking model made in 9.1. From 9.1 onward, each object in the system has it's own lock, which can be acquired in different modes - read, intent-to-write (IX,) etc. Often, when performing maintenance on one object, other related objects must be additionally locked. This chaining can appear to lock up the entire server. One common example is when performing dimension maintenance on a common/popular dimension, which in turn locks any cubes using that dimension, which in turn locks any cubes downstream from those cubes (via rules.) When those cubes are element attribute cubes, the corresponding dimension may need to be found and updated, and the cycle repeats.

That said, there are some known actions (or combinations of actions) to stay away from (e.g. mid-day dimension maintenance.) Debug logging and Ops Console can be helpful in finding and avoiding your specific pain points.