c1iscex floods the network with about 1 gigabyte of error messages in a few seconds, writing to a log file in /opt/rtcds/caltech/c1/target/fb/logs/
I commented out the following line in the rc.local file on the fb machine in the /diskless/root/etc/ directory:
#nice --20 ./mx_stream -s "$SYSTEMS" -d fb:0 >& logs/$HOSTNAME.log&
This disables the automatic start up of the mx_streams code on all the front ends. This will prevent the network being brought to its knees by c1iscex while we debug the problem.
It also means on a reboot of the front ends, the mx_stream process needs to be started by hand until this change is reverted.
To do this, log into the front end and then change directory to /opt/rtcds/caltech/c1/target/fb
For c1sus, run:
./mx_stream -s c1x02 c1sus c1mcs c1rms c1rfm -d fb:0
For c1ioo, run:
./mx_stream -s c1x03 c1ioo -d fb:0