Re: [Ssh-sftp-perl-users] Question to BatchMode

[Stefan A. <ste...@gm...>]

>From Russ Brewer
> Stefan,
>
> eval {
>    my ($stdout, $stderr, $exit) = $ssh->cmd(BashScript);
>
>     @sdtout_result = split(/\n/, $stdout);
>     @stderr_result = split(/\n/, $stderr);
>     $BashScript_exit_status = $exit;
> };
>
> if (($BashScript_exit_status ne 0) || ($@))
>    # then there was a problem
>
>      if ($@) {
>         print "\$@ = $@\n";
>      } else {
>         print "ERROR: remote script exit status =
> $BashScript_exit_status\n";
>      }
>
>      foreach (@stderr_result) {
>        print "$_\n";
>      }
>
> } else {
>
>     # No error was detected
>      foreach (@stdout_result) {
>         print "$_\n";
>      }
> }
>

The script idea itself works, but it does not solve my problem. I will
explain my problem again in detail at the bottom.


> I have not used the register_handler command because the above code gets me
> what I need.
> Be sure to program BashScript to exit with a non_zero value if it encounters
> an error while running on the remote server.
>
> You can also program BashScript to run STDERR error free when it encounters
> no run problem and then test @stderr_result for content. If content exists,
> something went wrong. The exception might be an FTP command which utilizes
> the STDERR channel for routine FTP communication. I like to write BashScript
> to log its own errors remotely and then have it check that log for size.  If
> the remote error log holds content then terminate BashScript with exit(1)
> instead of exit(0).
>
> I hope this helps. I works for me.
>
>
> Russ Brewer
>
>
>

Original Post
>>
>> I'm running Net::SSH::Perl 1.34.
>>
>> I execute a bash-script via ssh on a remote server.
>>
>> I register two handles for stdout and stderr:
>>
>>                $ssh->register_handler(
>>                        "stdout",
>>                        sub {
>>                                my ( $channel, $buffer ) = @_;
>>                                print $buffer->bytes;
>>                                print LOGFILE $buffer->bytes;
>>                        }
>>                );
>>                $ssh->register_handler(
>>                        "stderr",
>>                        sub {
>>                                my ( $channel, $buffer ) = @_;
>>                                print $buffer->bytes;
>>                                print LOGFILE $buffer->bytes;
>>                        }
>>                );
>>
>> Then is send my command with:
>>
>> $ssh->cmd("BashScript");
>>
>> wait until it finishes.
>>
>> The question I have is:
>>
>> How do I detect a connection loss?
>>
>> The script on the remote server my take a very long time without any
>> activity on the console.
>>
>> I have tried invoking my connection with "options => [
>> "ServerAliveInterval 15 ServerAliveCountMax 10" ]" or "options => [
>> "BatchMode yes" ]". However whenever I test it, by pulling out the
>> network cable while the script runs, it just sits there. The longest I
>> have waited is 45 min until I aborted manually.
>>
>> Shouldn't the connection be checked every 300 seconds with "BatchMode
>> yes"?
>>
>> Any help would be appreciated.
>>
>>
>> ------------------------------------------------------------------------------
>> Open Source Business Conference (OSBC), March 24-25, 2009, San Francisco,
>> CA
>> -OSBC tackles the biggest issue in open source: Open Sourcing the
>> Enterprise
>> -Strategies to boost innovation and cut costs with open source
>> participation
>> -Receive a $600 discount off the registration fee with the source code:
>> SFAD
>> http://p.sf.net/sfu/XcvMzF8H
>> _______________________________________________
>> Ssh-sftp-perl-users mailing list
>> Ssh...@li...
>> https://lists.sourceforge.net/lists/listinfo/ssh-sftp-perl-users
>
>

Okay:

I have two machines. One runs the perl script. The script connects
through ssh and runs a bash script on the second machine. This bash
script can run quiet some time without giving back any feed back, but
it still runs.

Here is the example how I would run it directly from the command line with ssh:

ssh -o "ServerAliveInterval 15" user@2ndMachine /home/user/BashScript.sh

One thing that is important for me, is to detect if I'm still
connected to the server. If I run it in the command line as described
and pull the network cable on the second machine, I get after some
time a time out. This is a desired behavior for me.

I have tried the same thing with a perl script. Heres an example script:

#!/usr/bin/perl -w

use strict;
use warnings;
use Net::SSH::Perl;

{
	my $ssh = Net::SSH::Perl->new( "2ndMachine", protocol => 2, options
=> [ "ServerAliveInterval 15"] );
	$ssh->login("user", "password");
	my @sdtout_result;
	my @stderr_result;
	my $BashScript_exit_status;
	eval {
	   	my ($stdout, $stderr, $exit) =
$ssh->cmd("/srv/import/bin/backup_topsis.sh");
	
	    @sdtout_result = split(/\n/, $stdout);
	    @stderr_result = split(/\n/, $stderr);
	    $BashScript_exit_status = $exit;
	};
	
	if (($BashScript_exit_status ne 0) || ($@))
	   # then there was a problem
	{
	     if ($@)
	     {
	        print "\$@ = $@\n";
	     }
	     else
	     {
	        print "ERROR: remote script exit status = $BashScript_exit_status\n";
	     }
	
	     foreach (@stderr_result)
	     {
	       print "$_\n";
	     }
	
	}
	else
	{
	
	    # No error was detected
	     foreach ( @sdtout_result )
	     {
	        print "$_\n";
	     }
	}
	
}

In this case, however, I do not get a time out if I pull the network
cable on the second machine. The script will just sit there forever.
It is important for me to detect network errors and in this case I can
not detect them.

I do not have a problem detecting the exit code of the remote script.
That is already possible. I have used the handler method (see first
post), because I wanted a continuous output of the progress of the
remote script.

If anyone has an idea how I can detect any network errors, it would be great.

Stefan