Page 1 of 1
Run in parallel, using data of another updatable cube and Wait
Posted: Tue Jun 25, 2024 9:03 am
by EP_explorer
I wrote a process1 which launch 6 processes in parallel (I'm using RushTI if it is important to launch parallel processes) which updates Cube1 and use during update data which calculated in Cube2.
Before launch process1 I need launch process2 which updates Cube2.
If I launch process2 independently of process1 - everything ok in Process1. 6 process launch and run without problems
If if launch process2 with process1 (it doesn't matter I put ExecuteProcess('process2') in process1 at start position or I create process3 with launch process1 and process2) I have problems:
RushTI run processes in parallel but they put by TM1 engine in status WAIT. It seems the main problem is that TM1 engine thinks that data of Cube2 isn't updated yet.
I tried to put command SaveDataAll after ExecuteProcess('process2') in process1 but it didn't help.
Of course it is possible to put one button for users for update Cube2 by process2 at first and after it start by click to another button process1, but may be you can give me another solution?
Re: Run in parallel, using data of another updatable cube and Wait
Posted: Tue Jun 25, 2024 1:00 pm
by ardi
EP_explorer wrote: ↑Tue Jun 25, 2024 9:03 am
I wrote a process1 which launch 6 processes in parallel (I'm using RushTI if it is important to launch parallel processes) which updates Cube1 and use during update data which calculated in Cube2.
Before launch process1 I need launch process2 which updates Cube2.
If I launch process2 independently of process1 - everything ok in Process1. 6 process launch and run without problems
If if launch process2 with process1 (it doesn't matter I put ExecuteProcess('process2') in process1 at start position or I create process3 with launch process1 and process2) I have problems:
RushTI run processes in parallel but they put by TM1 engine in status WAIT. It seems the main problem is that TM1 engine thinks that data of Cube2 isn't updated yet.
I tried to put command SaveDataAll after ExecuteProcess('process2') in process1 but it didn't help.
Of course it is possible to put one button for users for update Cube2 by process2 at first and after it start by click to another button process1, but may be you can give me another solution?
In the process where you create the TaskFile and launch RushTI, do you run the ExecuteCommand with a Wait Flag?
It could be that some of the subsets/views you create are causing the lock
Re: Run in parallel, using data of another updatable cube and Wait
Posted: Tue Jun 25, 2024 3:40 pm
by EP_explorer
Yes.
I create TaskFile in Prolog
and run ExecuteCommand with rushti in Epilog
With Status Wait
Re: Run in parallel, using data of another updatable cube and Wait
Posted: Wed Jun 26, 2024 2:09 am
by ardi
EP_explorer wrote: ↑Tue Jun 25, 2024 3:40 pm
Yes.
I create TaskFile in Prolog
and run ExecuteCommand with rushti in Epilog
With Status Wait
Do you create any Subset/View in the Prolog of the process? If these Views or subsets are referenced in the processes triggered by RushTI, then this might be the reason why your process goes on Wait.
Re: Run in parallel, using data of another updatable cube and Wait
Posted: Fri Jun 28, 2024 12:36 pm
by EP_explorer
ardi wrote: ↑Wed Jun 26, 2024 2:09 am
EP_explorer wrote: ↑Tue Jun 25, 2024 3:40 pm
Yes.
I create TaskFile in Prolog
and run ExecuteCommand with rushti in Epilog
With Status Wait
Do you create any Subset/View in the Prolog of the process? If these Views or subsets are referenced in the processes triggered by RushTI, then this might be the reason why your process goes on Wait.
You know it was quite strange. Subset and Views are created and deleted in 5 processes (in process2) which updates Cube2.
I commented processes one by one and found one which was the reason of the problem. But it was made like others. Completely no differencies
That helps - I use Sourse View in the problem process which doesn't create and delete views and subsets during the process but based on mdx subsets
so new views and subsets aren't created during the process.
Re: Run in parallel, using data of another updatable cube and Wait
Posted: Fri Jun 28, 2024 12:38 pm
by EP_explorer
.