If I were you, I would run all the 1000 commands as background jobs, just to avoid any network issues. You list the commands you want to run into a shell script, upload the shell script via ssh, then execute the 1000 command script as nohup background jobs. Then come back in at a later time, and collect the results.

Using Net::SSH2 it is necessary to use the following shell syntax to execute background commands. Notice the closing of the standard filehandles, and the redirects to log files as output.

If you did it correctly, it should avoid any hassle with the ssh connection.

my $chan = $ssh2->channel();
$chan->blocking(1);

$chan->exec("nohup /home/user/myscript > myscript.out 2> myscript.err 
+< /dev/null &");

$chan->send_eof;

exit;
[download]

This is largely hardware dependent, but on standard hardware spawning 1000 process will kill the performance on each one due to swapping. You are generally far better off spawning a number just less than the number of available processors/cores, and issuing more commands as the old ones complete.

But i am worried about run it with 1000 items and how the remote server reacts.

Probably the same way as your local computer would react if you ran those 1000 commands locally. Unless of course the machines are very different, in which case you should consult the admin of the remote machine, and ask him.

grep -E ' 900 ' | grep -v 1900 | grep -E -v '0 RE' | grep -E -v '0 WE' | grep -E -v '0 WAV' | grep -E -v '0 WA' | awk '{print \$3\";\", \$6\";\" }'");

That's an awful lot of greps, and they're all of the same "fixed-string"/-F variety. You could probably reduce the above to something like:

my @ls = $ssh2convey99->capture(q/disptest | fgrep ' 900 ' | fgrep -v 
+1900 | grep -vE '0 ([RW]E|WA)' | awk '{print $3";", $6";" }'/);
[download]

...or replace the whole grep-awk chain with a Perl script!

update: "0 WA" already takes care of "0 WAV"

Guys you are really quick. I changed the grep statement as written by "hazylife". Added saving to file. About the answers from "kennethk" und "zentara" i need to expand my horizon the next days. For testing it now with looping up to 20 times, i just added a sleep() (temporarely). The script will run after normal working hours of the server at night. So itīs not needed to be as quick as possible.

#!/usr/bin/perl -w
use Net::OpenSSH;
use File::Copy;
use Cwd;
use strict;
use warnings;

#connection:
  my $ssh2convey99 = Net::OpenSSH->new('...');
  $ssh2convey99->error and
    die "Couldn't establish SSH connection: ". $ssh2convey99->error;

#get some data from remote host and fill array    
  my @ls = $ssh2convey99->capture(q/disptest | fgrep ' 900 ' | fgrep -
+v +1900 | grep -vE '0 ([RW]E|WAV?)' | awk '{print $3";", $6";" }'/);
    $ssh2convey99->error and
        die "remote ls command failed: " . $ssh2convey99->error;
    
    
# include variables and arrays for data preparation
    my @split_array;
    my @matrix_array;
    my $mitzl = 0; 
    
#    prepare the data and store in the matrix
    foreach my $el (@ls){
        $mitzl ++; 
        @split_array = split(/;/,$el);
        $matrix_array[$mitzl] = [@split_array];
    }
    

#including variables, arrays, date and file for getting the needed dat
+a for all list elements
     
    my @bt_matrix_array;
    my $length_matrix_array = @matrix_array;    
    
    
    my ($sec,$min,$hour,$mday,$mon,$year,$wday,$yday,$isdst) = localti
+me(time);
    my $datum = sprintf "%04d%02d%02d" , $year+=1900 , $mon+=1 , $mday
+ ;
    my $file = "$datum.csv";

    #running thru the list and send command to remote host  for every 
+element.
    #capture everey command return in new array
     for(my $i = 2;$i <= 10;$i++) { 
        my @bt = $ssh2convey99->capture("bt $matrix_array[$i][0]");
        
        #kick out the first two lines because i donīt want them
         splice(@bt, 0, 2) ;

        #write the actual array in a file. make the file if not existi
+ng. if already existing -> put the new data below the old one. add so
+me meta infos
        open (DATEI, ">>$file") or die $!;
            print DATEI "TE: $matrix_array[$i][0]   Boxtype: $matrix_a
+rray[$i][1] \n @bt \n\n";
            #make a break 
            sleep(1);
        close (DATEI);
     }
     undef $ssh2convey99;


my $dir = cwd;
my $newdir = '//mnt/SharedFolder/';


move("$dir/$file", "$newdir/$file") or die "konnte Verzeichnis nicht o
+effnen: $!\n";

    ##http://perlmonks.org/index.pl?node=1080097
[download]

fgrep -v +1900

That should probably be fgrep -v 1900

Appendix: I get this error on every executed statement from the remote host
"tput: No value for $TERM and no -T specified"

somebody knows what to add on the ssh connection to get rid of it?

the script wil be used as backup procedure every day to collect a selected type of data for long time storing. It is running within the same internal network.

what i meaned to ask is: is it the right way to use the capture() function again and again. Will this function open always a new instance of the connection? Thanks, JGMonk