The 5112 Synchronization example won't work with the 5114 because the example tries to use the "Trigger to RTSI Delay" and "Trigger from RTSI Delay" attributes, which the 5114 doesn't support. The correct method to fully synchronize 5114 to high precision is to use TClk.
As you discovered, though, RIS mode is not supported while using TClk. I want to make sure you understand what is happening with RIS. RIS takes many records of a periodic signal, determines the phase difference between (since the trigger detection circuitry has sub-sample precision), and combines the records into a single record that is effectively oversampled. So, if the max real time sample rate of the 5114 is 250 MS/s, and you set a sample rate of 1 GS/s, the driver will acquire at least four records, and interleave the samples of the waveforms based on their relative phases. So you'll get four RIS samples for each real-time sample, or 4X oversamples (1 GS/s). This works as long as the waveform is periodic- you assume each record is effectively the same waveform so you can interleave different snapshots of time.
From this discussion, you can see that the points of the RIS waveform come from random points in time. So it's hard to talk about sychronization when using RIS. My guess is that you wanted a synchronized acquisition, then wanted an increased sample rate, so you set a higher rate. The driver assumed that you were trying to use RIS with TClk synchronization and returned an error.
So, my recommendations:
What sample rate does your application really require? The max real-time sample rate of the 5114 is 250 MS/s. You can TClk synchronize up to this rate.
If you want to use an RIS-like scheme and use TClk, I suggest configuring a multi-record acquisition using TClk, fetch the records back, and implement RIS in software.
Does any of this help?