To reproduce the 2.8 problems seen on H1 today, we reconfigured the /etc/init.d/mx_stream startup script to only use card0 10GE card on x1dc0. Front ends still have unique end-point identifiers (line number in rtsystab file), but are now not split over the two 10GE cards. Once this was done the DAQ data problem seen on H1 was reproduced on the DTS.