Strange behaviour of Subsets used in DBRW

[hju]

Hi all,
we are facing a strange behaviour of subsets used in DBRW in Perspectives (not PAX as mentioned here)

TM1-Version is PA2.0 (Product version 11.8.01000.6)

The application uses subsets (filters) defined by users to get sums and list of elements. Everything went fine so far for years.
- A single subset is created via TI with specified elements.
- The subset is accessible as pseudo consolidation (described here) in TM1Web (sum via DBRW) and CubeViewer and perspectives

Now we enhanced the feature to enable the user to create multiple subset (around 5 to 15) in on step (just a loop over the existing process)

Everything works as expected
- All subsets are created with the specified elements
- all subset are accessible as pseudo-consolidation in CubeViewer and TM1WEB retrieving the sum

BUT:
In Perspectives itself we encounter the following cases (independent of the number of subsets created in one loop)
- all subsets created in the loop are available via DBRW (works as expected)
- all subsets created in the loop show *KEY_ERR (but they exist and can be evaluated via CubeViewer or TM1Web as pseudo-consolidation)
There is no real pattern in this. All names are valid, no errors, all subsets have elements

ADDITIONAL:
After some hours (here observed: 2,5h), the subsets producing *KEY_ERR earlier are now working as expected in perspectives (no more *KEY_ERR)

We tried so far with no effekt:
- SaveDataAll
- Reload/Rebuild of sheet
- Closing perspectives worksheet and reopening

Anyone observed this before?
Any hints?

Sincerly
Holger

[gtonkin]

What happens if you create the new sets, close Perspectives and Excel then open and connect again?
Could be a meta data sync issue where new objects are not available/recognized.

[hju]

What happens if you create the new sets, close Perspectives and Excel then open and connect again?
Could be a meta data sync issue where new objects are not available/recognized.

Hi gtonkin,
my users checked it and i think you got it. After completely closing and reconnect all subsets can be adressed.
Indeed seems to be a sync issue.
Strange that it occures only sometimes...

Is there a way to to force a resync by ti (somthing like RefreshMdxHierarchy for cognos) ?
Maybe a client configuration issue in perspectives?

Thanks in advance

[gtonkin]

I have used M_Clear in macro code previously but has some additional erratic behaviour like suddenly having your connection close.

Hopefully someone else has a better alternative…

[burnstripe]

long shot, does the tm1p.ini refer to DimensionDownloadMaxSize? If so try removing it.

[hju]

Hi burnstripe,
the DimensionDownloadMaxSize parameter im tm1p.ini has no value assigned. Nevertheless the users removed the line and did a test.
Sadly there was no positve effect...

Any further ideas on how to solve this sync-issue are welcome...