Can anyone speak to the advisability on a go-live weekend of copying a development project (xml, cpf etc) into production and just using it there as distinct from starting with the production project and running the xml scripts saved in development (scripts created in manually upgrading the development project to accommodate a revised source database - note that the production database is to be upgraded on that same weekend first)?
This would save us a lot of time running the scripts, I think. But I want to check my understanding, hear any gotchas etc
Much appreciated
Thanks
If you need to accomodate changes within the source DB within such a short time-frame, would it not be safer to work in regression-tests on the 'old' and 'new' model to make sure everything will work out fine (delivering identical results) ? Assuming that you adjust the model layer to fit onto the changed datasource table definitions, that is. Just rolling out scripts and hope that this covers 100% seems a bit optimistic
We have a relatively small number of Cognos reports and don't have the time/staff to introduce an automated system of regression testing.
The database changes and model changes were delivered in tandem by a third party vendor.
Our customizations to that database and the model are small in comparison. People will be testing all of this in our development system
My question concerns migrating to production
is it better to
- take the model re-delivered by the vendor and apply our scripted re-customizations in production
or
- take the model re-delivered by the vendor plus the changes made in FM for our re-customizations in our development environment and just use that model in production
Thanks
Can't answer that. It sounds fairly academic. We never publish models to production. The packages within developement are imported into test environments , tested and then used with installation sets (packages+reports) for various production environments with customers