sunckell wrote:
Hello Perl People....
I was wondering if someone could enlighten me on an observation I
have been seeing..
A little Background info:
Where I work we have been seeing very slow ssh connections times
(it's not DNS. Maybe NIS, NFS, but definitely not DNS).
Anyways I wanted to write a script that would log the time it
takes to connect to a remote server via ssh and issue a command (uname
-n). The tricky part was that I wanted the remote connections to kick
off every 60 secs. (some ssh connections take a lot longer. 120+ in
some cases). The log should enable me to pinpoint any time
correlation to the long ssh connect occurrences (every 5 minutes? 1/2
past every hour? etc).
So here we are back to the script.. The script works as
expected, with one item of strangeness. I was expecting the log
messages to be printed in the order they complete. For example, if the
connection attempt at 12:01 took 75 seconds to return, and the
connection attempt at 12:02 to 5 seconds to return, I would expect the
12:02 message to appear first then the 12:01, but it doesn't. The
12:02 connection won't print until the 12:01 returns.
Here is my code:
use strict;
use Getopt::Std;
use Sys::Hostname;
use File::Basename;
use Net::SSH qw( sshopen2 );
use Time::HiRes qw( gettimeofday tv_interval );
$|++;
$SIG{CHLD} = 'IGNORE';
my %opts;
getopts('dr:', \%opts);
my $localhost = hostname();
my $r_host = $opts{'r'} ? $opts{'r'} : die "need remote host<-r>
\n";
That is usually written as:
my $r_host = $opts{ r } or die "need remote host<-r>\n";
my $debug = $opts{'d'};
my $short_name = basename("$0", ".pl");
# --- set the counter to zero. We'll print the counter to the log for
easier sorting
# --- of data where the ssh connections take longer than a minute to
return.
my $ct = 0;
while (1) {
my $pid = fork();
if ($pid) {
# parent
} elsif ($pid == 0) {
# child
my $p = _initialize($ct);
waitpid($p, 0);
exit 0;
} else {
die "couldnt fork: $!\n";
}
That won't work correctly if fork returns undef and sets $! because
undef in numerical context is the same as 0.
$ perl -le'print "undef is ", undef == 0 ? "" : "NOT ", "equal to 0"'
undef is equal to 0
And without warnings enabled you won't catch this mistake.
See:
perldoc perlipc
For how to verify that fork() worked correctly.
$ct++;
sleep 60;
}
# ---
# --- sub: initialize
# --- descr: for easier viewing of the code for the fork process.
# --- returns: nothing
# ---
sub _initialize{
my($count) = @_;
my $now = scalar localtime();
"my $now = " forces scalar context on the right hand side of the
assignment so the use of "scalar" is redundant.
my $start = [gettimeofday()];
my $ret = connect_to_remote_host("$r_host");
my $elapsed = tv_interval( $start );
if ($count< 10){
$count = "0$count";
}
if ($count< 99){
$count = "0$count";
}
That is usually performed with sprintf:
$count = sprintf '%03d', $_[ 0 ];
my $msg = "$count - [$now] $localhost to $r_host Elapsed time:
$elapsed ";
write_to_log($msg);
return;
}
John
--
Any intelligent fool can make things bigger and
more complex... It takes a touch of genius -
and a lot of courage to move in the opposite
direction. -- Albert Einstein
--
To unsubscribe, e-mail: beginners-unsubscr...@perl.org
For additional commands, e-mail: beginners-h...@perl.org
http://learn.perl.org/
