.
.
my $ssh2 = Net::SSH2->new();

$ssh2->connect('xxx.xxx.xxx.xxx') or die;

$ssh2->auth_password('user','pass')
  or die "Unable to login $! \n";

my $chan1 = $ssh2->channel();

$chan1->blocking(1);

$chan1->exec('echo $PATH');
$chan1->read($buf, $buflen);
print "BUF: $buf\n";

$chan1->close;
.
.
[download]

I've messed around a lot with blocking and non-blocking connections without much luck, but I'll give your code segment a go.

Thanks, R

use Net::OpenSSH;

my $ssh = Net::OpenSSH->new($remote_host, user => $remote_user, passwo
+rd => $password);
$ssh->error and die "Unable to connect to $remote_host: " . $ssh->erro
+r;

my $fh = $ssh->pipe_out('ls -la /Users/vortura')
  or die "unable to run command: " . $ssh->error;

while (<$fh>) {
  print "OUT: $_";
}

close $fh or die "close failed: $?";
[download]

I hadn't seen that module. I will give it a try. I see it does it's business by running the openssh binary as well, but the multiplexing approach looks a lot nicer than the mess I've inherited.

Thanks!

BTW, this is how you read from both stderr and stdout:

use Net::OpenSSH;

my $ssh = Net::OpenSSH->new($host, ...);
$ssh->error and die "connection failed: " . $ssh->error;

my (undef, $ofh, $efh, $pid) = $ssh->open_ex({ stdout_pipe => 1,
                                               stderr_pipe => 1 },
                                             @cmd)
    or die "ssh command failed: " . $ssh->error;

while(1) {
  # select loop to read from $ofh and $efh without blocking
  # ...
}

waitpid($pid, 0); 
print "result: $?\n";
[download]

#!/usr/bin/perl
use warnings;
use strict;
use Net::SSH2;
use Data::Dumper;

# assuming a user named 'z' for demonstration 
# connecting to localhost, so you need your sshd running

# see maillist archives at
# http://lists.sourceforge.net/lists/listinfo/ssh-sftp-perl-users
# for deeper discussions

my $ssh2 = Net::SSH2->new();
$ssh2->connect('localhost') or die "Unable to connect Host $@ \n";
$ssh2->auth_password('z','ztester') or die "Unable to login $@ \n";

#shell use
my $chan = $ssh2->channel();
$chan->blocking(0);
$chan->shell();
print $chan "ls -la\n";
print "LINE : $_" while <$chan>;

print $chan "who\n";
print "LINE : $_" while <$chan>;

print $chan "date\n";
print "LINE : $_" while <$chan>;


$chan->close;
__END__
[download]

$chan->shell();
print $chan "ls -la\n";
print "LINE : $_" while <$chan>;

print $chan "who\n";
print "LINE : $_" while <$chan>;

print $chan "date\n";
print "LINE : $_" while <$chan>;
[download]

That is not reliable!

The remote shell doesn't signal in any way when the commands run through it exit to its parent process, the ssh server. If the remote command does not run fast enough, the ... while(<$chan>) loop will finish without having read the full output, and the remaining data will be delayed until the next loop (or even later).

Net::SSH::Expect that uses a similar approach, solves that problem examining the remote output, looking for the shell prompt in order to detect when a subcommand has finished.

Indeed, I've had more success with shell, but as salva points out, occasionally I end up with partial output. The other thing is that it's hard to distinguish stderr from stdout when running a command under shell.

It seems like Net::SSH2 has a new maintainer, so hopefully that module will start to see some improvement soon. In the meantime, it looks like Net::OpenSSH is the way to go for me.

R