Using named pipes

By Greg No comments

We run Analysis Services on the same server as an asynchronous connection-string-d-lockcopy of  SQL (mainly for licensing reasons). We have been having problems when processing the cube that it drops the TCP connection to
SQL. It was suggested that we use Named Pipes instead of TCP connections. We don’t want to turn off TCP completely, so how do you tell a connection to use named pipes?

Turns out it’s pretty easy. In the connection string we just need to add a np: to the start of the server name. So to connect to your local database, your connection string looks something like this:

Server=np:(local); Database=DatabaseName; Trusted_Connection=True; Encrypt=True; Connection Timeout=30;

 

Leave a Reply