"Stuart Russell" <stuartr@easter.euro.csg.mot.com> said:
Stuart> I am new to the scotty program and environment ( like 1 day )
Stuart> and would like to know if it is possible through the toolkit
Stuart> to check if a specific port on a server is being used , and by
Stuart> what.
Stuart> I have daemons which I would like to check are on the server
Stuart> without sending messages to the daemon process.
There is a difference between checking if a port is `in use' and
making sure that there is a server that will answer requests send to
the port. Checking if a port is `in use' can be done by sending SNMP
queries to dump the tcpConnTable or the udpTable. In Tkined, you can
use the following commands (see snmp-trouble.tcl for the
implementation):
SNMP-Trouble->TCP->TCP Connections
SNMP-Trouble->UDP->UDP Listener
Checking whether a service is actually available requires to probe the
server. You can try to connect to a TCP based server, but this still
does not tell that there is actually any service - but it is usually a
bit more than just checking the port status. Sun RPC services can be
probed by calling the NULL procedure. In Tkined, you can use the
following commands (see ip-trouble.tcl for the implementation):
IP-Trouble->TCP Services
IP-Trouble->RPC Services
Probing a UDP based server requires to send a UDP packet that makes
sense to the server. This is not possible without coding it for every
service individually.
Finding out what kind of service is available on a port that is not a
well-know port is not possible without having access to additional
information. There are proposals for MIBs that will give you more
information, e.g. the application MIB defined in RFC 1565. I do not
know if there are any agents that implemented this MIB.
Juergen