CounterReader exception: "This is not valid for acquisitions without a buffer."

[saratoga2]

I am trying to trouble shoot a problem with counter reading that occurs only under some circumstances.  If a certain sequence of actions happens in my software, running a multisample read on a CounterReader returns the following error:

By setting Number of Samples per Channel to -1, you indicated that all available data should be read. This is not valid for acquisitions without a buffer.

Specify a value greater than or equal to zero for Number of Samples per Channel. Do not specify a value of zero for Buffer Size when configuring the input buffer.

 My code is:

myXCounterTask = gcnew Task("X Counter");

myXCounterTask->CIChannels->CreateLinearEncoderChannel (..)

myXCounterTask->Timing->ConfigureSampleClock("/Dev1/PFI0", 8000, SampleClockActiveEdge::Rising, SampleQuantityMode::ContinuousSamples, numberOfSamples );

myXCounterTask->Stream->Buffer->InputBufferSize = numberOfSamples * 20;

counterXInReader = gcnew CounterReader(myXCounterTask->Stream);

array<double>^ posX = counterXInReader->ReadMultiSampleDouble(-1);

What is confusing is that this code works fine generally, but fails if certain sequences daqmx commands are run previously, which makes me think that I am not initializing something correctly and depending on it to be in a default state.  I'm trying to understand which buffer the error is referring to, and how I can check it's size?  It is apparently not the Task's buffer, since that is set.  Are there other buffers associated with the CounterReader?

[saratoga2]

After a lot of further testing, the problem had to do with the order in which different hardware devices in my system were being initialized.  For reasons I do not yet fully understanding, doing it in a certain order seemed to reset part of the DAQ state, which is probably why I got the confusing error message.

I will say though that some documentation would have made this a lot easier to trouble shoot.  There is so little information provided by NI that if you google the error message, this thread is now the #1 result.