Can one interface UFL write two independant servers ?
If it's possible, wath are best practices ?
The recommended method would be to run two instances of the PI UFL interface (you can do this on the same machine), each configured to send data through to their respective independent PI Data Archive.
You would have two PI UFL interface windows services functioning independently (able to start and stop the operation of one without effecting the other).
The configuration is simply repeating the process of creating the PI UFL interface again after completing configuration of the first, but for the second instance select the second PI Data Archive node for the host PI Data Archive (data sink) and a new .ini configuration file. PI UFL guide
An alternative approach (not recommended unless you are running in a non-Windows platform or the PI Server receiving the data is older than version 3.4.375) is possible if you have the exact same data going into the same PI Tags on your two independent PI Data Archives. This approach leverages PI Buffer Server (API buffering) and its support of n-way buffering to independent PI Data Archives to send the data more details on its configuration: here
thanks for informations.
can the two PI UFL interface have the same interface ID, or interface ID must be diffrente ?
I think the interface ID can be the same. If the servers are indenpendant, then the ID is not needed to identify between the two. The ID is used to identify which interface a PI Point belongs to if there are multiple instances of the same interface writing to the same PI Data Archive.
Making your point source unique is a best practice: KB00296 - Best Practice: Assigning a unique point source for each instance of an interface
In this case, the servers are independent, so not necessarily applicable, but wanted to mention it for future consideration.
Retrieving data ...