I'm trying to deploy a simple shiny app which has a mssql connection on to RStudio Connect. In RServer i'm logged-in with ldap. Running the app in RServer is working fine. When i'm trying to deploy the app i'm getting the following in RStudio Connect.
2020/10/26 19:04:50.881006533 Error in value[[3L]](cond) :
2020/10/26 19:04:50.881026964 nanodbc/nanodbc.cpp:983: 00000: [unixODBC][Microsoft][ODBC Driver 17 for SQL Server]SSPI Provider: No Kerberos credentials available (default cache: KEYRING:persistent:994) [unixODBC][Microsoft][ODBC Driver 17 for SQL Server]Cannot generate SSPI context [unixODBC][Microsoft][ODBC Driver 17 for SQL Server]Invalid connection string attribute
2020/10/26 19:04:50.881096060 Calls: local ... tryCatch -> tryCatchList -> tryCatchOne -> <Anonymous>
2020/10/26 19:04:50.881099439 Execution halted```
ps. my authentication on RStudio connect is also via LDAP
LDAP authentication can mean different things. With RStudio Server, your Linux system is connected to LDAP, probably via sssd, and takes care of creating a Kerberos ticket during login. Since the R process executes with your user rights, it can use the Kerberos ticket for connecting with the database.
RStudio Connect connects directly to the LDAP server to authenticate users. These users (normally) don't exist on the underlying Linux system. And the R process supporting your Shiny app is run using the user rights of some technical user.
As for solutions, you have to first answer one question: When two different users access the app, should the data returned from the database be the same or should it be different, since these two users have different access rights on the DB? In other words, does the data access need to use the user's credentials or can it use the credential for some technical account?