Hi ​@JoDe,

I’m from the R&D Data and Analytics team.
May I know which version of self hosted BI is you are currently using (sql server based or datalake based)?
What are the pain points of that solution you are experiencing now?

Fabric open-mirroring is not a concept we are using or have tested yet.

Best regards,
Subash
 

Hi ​@subashfestus, We've been on IFS cloud since the 21R2, and we've not had great experiences with the self hosted BI setup. We've had: 

• multiple times where the environment settings suddenly were removed after updates
• Tables in the SQL environment suddenly containing duplicate records
• Data loads suddenly breaking because of sudden incorrect data types
• The setup for the tabular models and dataloads not being very straight forward and user friendly

Currently I believe we use both SQL server and datalake? But preferably we'd only write it to the datalake.

echo ​@JoDe 

​@PhillBronson, did you want to ask if this was fixed? 

It's not yet, although I will have conversations soon to hopefully get SOME movement in this.

Hi ​@JoDe and ​@PhillBronson,

1. This issue was mainly due to pipeline issues in IFS Managed cloud environments. So we deprecated the use of SNow form to submit parameters and now the parameters are captured through IFS Cloud web client and saved in the database itself. So this missing values issue should not occur now.
2. It’s bit unusual to have duplicate records in the SQL server side tables. When we create tables we usually create a PK on the ID column and with that we avoid creating duplicate records on the table. If you see duplicate records in the data load process it might be due to Oracle source containing duplicates.
3. We have seen this error with number columns with lengthy decimals. This is due to a limitation on SQL server side compared to Oracle. As a solution we round the numbers when we create the access views and the length issue should not happen on IFS data sources. Are these some kind of IAL based data sources or are you getting some other data type related error?
4. We have mainly 4 commands for db object deployment, tabular model deployment, data load and model processing. Can you be more specific and share some info about where it is problematic so we can try to improve the UX. The process was greatly improved compared to apps10 where doing a modification to a data source was very technical involving SSIS package changes as well.

Thanks and regards,

Subash 

​@subashfestus, I feel like we're going a little off topic here. 

The question here was if IFS supports Fabric

Thanks for the information ​@subashfestus 

Hi ​@JoDe,

We have the “Analysis Model - Power BI” solution and in that we utilize PBI workspaces connected to a fabric capacity. It is possible to use the solution with PBI embed capacity as well.

Other than that we don’t utilize any Fabric features in our solutions.

We haven’t done any research or tests with the open mirroring feature you mentioned.

Is this a IFS managed environment or a customer self hosted environment?

With a quick search I saw the following requirements to enable mirroring for an Oracle DB.

• LogMiner enabled – for tracking changes.
• Archive log mode – to store redo logs.
• Supplemental logging – for capturing detailed changes.
• On-Premises Data Gateway – to connect your Oracle server to Fabric.

To the best of my knowledge we don’t have supplemental logging enabled in our databases.

If this is a IFS Managed environment, then I don’t think the support for supplemental logging and on-premises data gateways are created by IFS cloud services as standard solution offerings. Better to check that with cloud service to get a confirmation.

But if this is a self hosted one, I think you should be able to do the necessary configuration changes on the Oracle database to get this working.

This is i’m assuming that technically this should be possible. Proper research and tests are needed. Also check whether it is allowed to enable these configurations on the DB level even on a self hosted scenario.

Best regards

Subash