TI processes that load a reporting cube loads to incorrect server

Post Reply
njcube
Posts: 4
Joined: Tue Apr 21, 2020 9:36 pm
OLAP Product: Architect
Version: TM1 11.0.6
Excel Version: Excel 2013

TI processes that load a reporting cube loads to incorrect server

Post by njcube »

Hello,

We have a chore that consists of multiple TI processes that loads a TM1 cube that we use for reporting. When we run this chore in our DEV server, it is updating the reporting cube in our PROD environment with our DEV data.

We can't find where in the DEV server TM1 install it is referencing the PROD server.

The TM1S.Cfg file designates the DEV server as the ServerName and AdminHost. The port number is different then the one on the TM1S.Cfg in Prod.

Where else can this connection be made?

Thank you,
Rich
declanr
MVP
Posts: 1815
Joined: Mon Dec 05, 2011 11:51 am
OLAP Product: Cognos TM1
Version: PA2.0 and most of the old ones
Excel Version: All of em
Location: Manchester, United Kingdom
Contact:

Re: TI processes that load a reporting cube loads to incorrect server

Post by declanr »

In a "standard" set up (if such a thing exists) it is not possible for a TI process to load data into a cube that is not in the same TM1 service that the TI process is running in (e.g. a CellPutN can only update a cube on the service in which the TI is executed.)

That being said there are ways you can run a TI on a different service e.g. TM1RunTI can be used to kick off a TI on Server B from Server A.
It would be unlikely that anyone would ever intentionally have a dev server run a process on a prod server BUT it is feasible that there is a master process on the Prod server that runs sub process (also on the Prod server) in parallel (using TM1RunTI) to make the load quicker.
If those references were hardcoded and copied back from Prod to Dev, the master process in dev could run the sub processes in Prod.

Its also possible that the Dev server has a process that exports data to a CSV file (or DB table) and then a second process imports data from that CSV file (or DB table.) Again, if hardcoded - both Dev and Prod could be updating the same file/table - which could cause Dev data to "leak" into Prod and vice versa.

The above are a couple of unlikely but possible culprits.
To find your culprit, I would start by running a transaction log on the Production cube and pinpoint when a cell was updated with Dev data. From that timestamp, check the server logs (for the prod server) to see if a TI process was running at that time. You can then narrow it down to the process that loaded the data.

Note it could also be possible from PAX or Perspectives (or a websheet) to move data between servers. But since you went down the chore route as a culprit I would focus on TIs as the likely cause first.


Also out of interest - you say that the chore in Dev updated the cube in Prod... did it ALSO update the cube in Dev? Or did the cube in Dev get missed altogether.
Declan Rodger
User avatar
Steve Rowe
Site Admin
Posts: 2410
Joined: Wed May 14, 2008 4:25 pm
OLAP Product: TM1
Version: TM1 v6,v7,v8,v9,v10,v11+PAW
Excel Version: Nearly all of them

Re: TI processes that load a reporting cube loads to incorrect server

Post by Steve Rowe »

You've not mentioned the source for the TI processes.

I'm guessing it is a SQL query.

Have you checked that any ODBC connections in the production environment are not pointing at a Dev environment or a cube of reference information has not been promoted with the Dev information in it?

It is far more likely that a TI in production is pointing at a Dev datasource than a TI in Dev is writing to a production cube. As Declan says, it is very very very unlikely to be running like that and not something could happen by accident.

How have you proved it is the Dev server updating Prod?
Technical Director
www.infocat.co.uk
njcube
Posts: 4
Joined: Tue Apr 21, 2020 9:36 pm
OLAP Product: Architect
Version: TM1 11.0.6
Excel Version: Excel 2013

Re: TI processes that load a reporting cube loads to incorrect server

Post by njcube »

Hello,

You are both correct! Thank you so much.

I don't know how I didn't notice it earlier but we have two chores running on the DEV TM1 instance. One is to update the reporting cube, the other is an export to SQL server. In the Export TI processes there are ODBC connections being made to the SQL Server database that PROD also utilizes. That is why they would override each other.

We do have another issue in the DEV TM1 environment that made us think that there was something more going on, but I'll make another post for that since it is not related.

Thanks again!
Rich
Post Reply