	C3 version 3.1.2:   Cluster Command & Control Suite
          Oak Ridge National Laboratory, Oak Ridge, TN,
    Authors: M.Brim, R.Flanery, G.A.Geist, B.Luethke, S.L.Scott
	        (C) 2001 All Rights Reserved



1: When specifying indirect remote clusters the default 
cluster on the remote machine is executed. If that machines
default cluster is the calling machine then there is a possibility
of an infinite circle. In other words each machines default cluster
is the other machine.

2: ckill reads the output of ps to retrieve the process ID. If
the version of ps does not return its output in the same format
as the ps included with redhat then it will fail.

3:  when starting a daemon or running an infinite process as follows:

cexec 'a.out &'

cexec will not return. This is a problem with ssh not closing and
C3 waiting for the ssh process to finish. If you ^C the cexec 
process after all the daemons/processes are started they will continue 
to run (use cexec 'ps -A |grep -i process_name' to check if you want)

4: C3 automatically detects if a cluster is local or remote by comparing
the current machines hostname to both the internal and external IP address
supplied for the cluster. In cases such that you are trying to execute on 
a remote cluster that has the same interface name as a local cluster, C3
will execute the command on the local cluster. For example, the following
machine's configuration file has this problem:

cluster torc { #local cluster
	torc:node0
	node1[1-4]
}

cluster xtorc { #direct remote
	xtorc:node0
	node[1-10]
}

Since torc's head-node must have the internal address as it's hostname "node0"
will match both clusters as local. For now, the only fix is to use and indirect
remote cluster.

if a bug is found please report it to:
luethkeb@ornl.gov
scottsl@ornl.gov
