Arlin and Caitlin - Thanks for the insights about the triplets. I did some experimentation with a test program and got it straightened out. Turned out that we had two problems to consider. First was the endian issue, which I expected but couldn't figure out why the byte swapping wasn't fixing it. Then I realized that the new box was using DAPL 2.0 and the others are 1.2 version. Redefining the message we are sending to account for both DAT_RMR_TRIPLET types took care of everything.
Chuck
_______________________________________________ general mailing list [email protected] http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
