You say that they are not running, but the errors you show indicate
that you cannot connect to them. So are they running?
If you look at all of the processes on the central manager (where
they should be running) do you see the collector and negotiator
running? Are there any error messages in the log files that they
produce?
They are not running:
OK, what is DAEMON_LIST defined to be on that host? It should list the
collector and negotiator.
1) If they aren't listed, add them. Just add COLLECTOR and NEGOTIATOR to
the list, separated with commas. Given that you already have the master,
startd, and schedd, it would look like this:
DAEMON_LIST = MASTER, COLLECTOR, NEGOTIATOR, STARTD, SCHEDD
2) If they are listed but you don't see them running, then they crashed.
Look for a CollectorLog and a NegotiatorLog and see if they tell us
anything good. Also look for core files in the log directory.