TM1 Forum

We have a situation repeating for last few days, where in a TI process that exports data from a cube into SQL table, is unable to load any rows - when executed as a chore. However, when run independently it is able to load the table with desired number of rows!

Looking at the log files, something that jumped out was that SaveDataAll was executed (by a separate chore) about a sec or 2 later. It took precedence, completed the work and then the export process finished its job.

This TI Process:

1. Reads data from cube and puts it into a file
2. Truncates the table
3. Inserts rows into the table from file

1st step takes about 140sec, 2nd one about 5s and 3rd one about 40s. So if you look below for the time it took to execute, it sounds about right.
So far I have tried to:

1. Run the TI process with my id multiple times and it works
2. Schedule it as a separate chore and it works!

If we can have SaveDataAll to begin about 10 min late, I think it will work out just fine. I'd like to know, does something happen during SaveDataAll that causes data to be not present in the cube for a brief period of time? What am I missing here?

I'd appreciate your input. Thanks.

I am sure it is puzzling to everyone, since there are lot of views with no replies.

Anyway, we altered the chore schedules and now SaveDataAll does not interfere with the export process. Table is getting populated with desired number of rows now!

This is something pretty basic within TM1. Save Data locks the service. If you had done a search on here you'd have found your answer pretty quickly.

Here's the sequence of events:

* Export process runs for 62s (during this entire time, it is writing to a file using AsciiOutput)
* SaveDataAll kicks off, runs for 0.14s and finishes it
* Export process continues to run for an additional 125s and finishes its job

Save Data locks the service

Are you saying that when it locks the service, it is somehow able to wipe out the contents of the file written in the 1st 62s before it was kicked off?

When it does lock the service during its execution, I am assuming it releases the lock after its execution. Not able to read a cube during the service lock ... I can understand it. So after 0.14s of lock, shouldn't the export process have access to the cube, to read from it and write to the file?

Please help me understand this ... I'd appreciate that. Thanks.

You didn't mention anything about data loss in your original post. Have you had any? I theory the export should start and the save will kick in. As an export is only a read it will be put on hold while the save completes. You really need to look at the admin guide which will fill in the blanks you have regarding chores.