RExecClient implements the rexec() facility that first appeared in
4.2BSD Unix. This class will probably only be of use for connecting
to Unix systems and only when the rexecd daemon is configured to run,
which is a rarity these days because of the security risks involved.
However, rexec() can be very useful for performing administrative tasks
on a network behind a firewall.
As with virtually all of the client classes in org.apache.commons.net, this
class derives from SocketClient, inheriting its connection methods.
The way to use RExecClient is to first connect
to the server, call the
rexec() method, and then
fetch the connection's input, output, and optionally error streams.
Interaction with the remote command is controlled entirely through the
I/O streams. Once you have finished processing the streams, you should
invoke
disconnect() to clean up properly.
By default the standard output and standard error streams of the
remote process are transmitted over the same connection, readable
from the input stream returned by
getInputStream() . However, it is
possible to tell the rexecd daemon to return the standard error
stream over a separate connection, readable from the input stream
returned by
getErrorStream() . You
can specify that a separate connection should be created for standard
error by setting the boolean
separateErrorStream
parameter of
rexec() to
true
.
The standard input of the remote process can be written to through
the output stream returned by
getOutputSream() .
_createErrorStream
(package private) InputStream _createErrorStream()
throws IOException
disconnect
public void disconnect()
throws IOException
Disconnects from the server, closing all associated open sockets and
streams.
- disconnect in interface SocketClient
getErrorStream
public InputStream getErrorStream()
Returns the InputStream from which the standard error of the remote
process can be read if a separate error stream is requested from
the server. Otherwise, null will be returned. The error stream
will only be set after a successful rexec() invocation.
- The InputStream from which the standard error of the remote
process can be read if a separate error stream is requested from
the server. Otherwise, null will be returned.
getInputStream
public InputStream getInputStream()
Returns the InputStream from which the standard outputof the remote
process can be read. The input stream will only be set after a
successful rexec() invocation.
- The InputStream from which the standard output of the remote
process can be read.
getOutputStream
public OutputStream getOutputStream()
Returns the OutputStream through which the standard input of the remote
process can be written. The output stream will only be set after a
successful rexec() invocation.
- The OutputStream through which the standard input of the remote
process can be written.
isRemoteVerificationEnabled
public final boolean isRemoteVerificationEnabled()
Return whether or not verification of the remote host providing a
separate error stream is enabled. The default behavior is for
verification to be enabled.
- True if verification is enabled, false if not.
rexec
public void rexec(String username,
String password,
String command)
throws IOException
Same as rexec(username, password, command, false);
rexec
public void rexec(String username,
String password,
String command,
boolean separateErrorStream)
throws IOException
Remotely executes a command through the rexecd daemon on the server
to which the RExecClient is connected. After calling this method,
you may interact with the remote process through its standard input,
output, and error streams. You will typically be able to detect
the termination of the remote process after reaching end of file
on its standard output (accessible through
getInputStream() . Disconnecting
from the server or closing the process streams before reaching
end of file will not necessarily terminate the remote process.
If a separate error stream is requested, the remote server will
connect to a local socket opened by RExecClient, providing an
independent stream through which standard error will be transmitted.
RExecClient will do a simple security check when it accepts a
connection for this error stream. If the connection does not originate
from the remote server, an IOException will be thrown. This serves as
a simple protection against possible hijacking of the error stream by
an attacker monitoring the rexec() negotiation. You may disable this
behavior with
setRemoteVerificationEnabled().
username
- The account name on the server through which to execute
the command.password
- The plain text password of the user account.command
- The command, including any arguments, to execute.separateErrorStream
- True if you would like the standard error
to be transmitted through a different stream than standard output.
False if not.
setRemoteVerificationEnabled
public final void setRemoteVerificationEnabled(boolean enable)
Enable or disable verification that the remote host connecting to
create a separate error stream is the same as the host to which
the standard out stream is connected. The default is for verification
to be enabled. You may set this value at any time, whether the
client is currently connected or not.
enable
- True to enable verification, false to disable verification.