Hi ,
Sorry if there is a previous related topic - I couldn't find: what is the best way of managing FM models between Test-UAT-Live?
Difference will be datasources and packages, in a simple save as these need to be updated each time.
Regards
You do not publish or even install Framework Manager on Live or UAT. Install only on the Test, publish the package on Test, transfer it to UAT and then LIVE using the deployment process. The Data Sources can have the same name on all three servers but will be pointed to different databases. As long as the Data Sources have the same name, the deployment would work properly. Whe you deploy the packages, do NOT include Data Sources and signons otherwise the Production Data Sources will start pointing back to Test.
Thanks for the reply but we have a shared application server so all 3 FM models are on the same server with different datasources, different packages but with the same objects.
So the question is what is the best way to carry the changes(to objects) from one model to another without changing the datasource & package names.
Regards...
One datasource can have multiple connections, making it possible to connect to different databases (you will be prompted to choose a connection).
But.. Cognostechie is right that on instance of FM is the way to go. In your case you willl need to synchronize models forever (for no good cause)
Did you mean you have 3 FM Models on the same server to publish the packages to Test, UAT and Live? 3 Models to publish to 3 servers?
Even if so, how is it publishing the packages to different servers? The publishing of packages does not depend on where the data sources are pointed to but where the Cognos configuration is pointed to. So those 3 FM Models would still publish the packages on only one server even if the data sources are using different connections.
I am aware of the best practise; thanks but this is the case. There is one app server and one gateway but 3 FM models (Test, UAT, Live) with different datasources and different packages.
Each time one of them is saved on the other (save as) and package name-datasource is updated.
I wonder if there is a easier way for this case; maybe synch or linked models ?
Thanks for the help anyway...
Regards
If you have kept all logs, then FM allows for running scripts from the transaction history. Each session is stored as a unit and allows you to pick what you need. So, theoretically you could play all actions you performed on 1 model against other models by using the log files..
I think the original poster's issue is that because his client has just one server on which all 3 environments are hosted (very strange case indeed), he has to maintain 3 different models because the data source name is different and since the model uses the data source name, he's sort of stuck with maintaining 3 models.
One way I can think of to reduce maintainence a little bit is that you make all your changes on your dev model (import new tables, add in calculations..whatever u need to do) and then use this as a base model to import into the other 2 FM models...while importing in the other 2 FM models, just pick IBM Cognos 8 Model, open up the 1st model and import the newly created/modified etc. query subjects in. This way, you don't need to replicate changes across the other 2 models.
Another way out would be to use the same datasource but different database connections within that datasource as suggested by cognostechie. The annoyance of database selection prompt (which btw can be a caching issue if users want to test against uat and prod databases simultaneously) can be addressed by creating 3 security groups - one for dev, one for uat and one for prod and adding these groups i the appropriate database connection so Cognos automatically knows which person shouldbe given which database.
But, this might not work completely in your case because I suspect that almost all testers are also users so there will be a large intersection and these intersection class won't be able to avoid the prompt between at least uat and prod (if not dev) selection.
Thanks Arsenal; that's the exact case I'm afraid.
I'll try the import option to see if it can reduce the maintainence or not compare to "Save as"
I'll take a look to synch as well.
Thanks all.
Regards,
You can't really do a "Save As" in your case because remember your Dev model, for example, has a different data source name so you can't just make changes there and then Save As also your UAT model - that will just end up duplicating your Dev model
Hi,
We also had the same requirement of the same sever for DEV, UAT and SIT. The datasources and the packages are different. What we have done is , created 3 separate CPF files and pointed the datasources to their respective, published package and have created separate folders for Dev, UAT and SIT in the portal and put the same reports in them. I guess this could be correct way if we have same server for all (dev, uat, sit)
Regards
Nag31