On Sun, Mar 8, 2009 at 21:41, Owen <[email protected]> wrote:
>>
>> I'd like to monitor a running script and if the script stops running,
>> restart it and continue to monitor it. I need a little help from the
>> community on the ins-and-outs of the details. But, basically I need
>> someone to look at the following code/pseudo-code and give
>> suggestions.
>>
>> 1) Check a script running.
>> 2) If it's running, sleep for 30 seconds then check again.
>> 3) If it's not running, restart it and continue to check after that.
>>
>> This is on a Red Hat box, so the first thing would be something like:
>>
>> While (1) {
>> my $process = `ps -ef | grep <process name> | grep -v grep`;
>>
>> if ($process) {
>> sleep 30} else {
>> exec (./<process name>) or print STDERR "couldn't exec
>> <process name>: $!";
>> }
>> }
>
>
> WARNING: Totally untested, but you will get the idea
>
>
> #!/usr/bin/perl -w
>
> use strict;
>
> my $program = "<process_name>";
> my $status = `/bin/ps cat | /bin/grep $program`;
>
> if ( length($status) > 0 ) {
>
> sleep 30;
>
> }
> else { exec "the_process" } # start program
snip
You need a fork before the exec or you will lose the monitor program
the first time it restarts the monitored processed:
#!/usr/bin/perl
use warnings;
use strict;
my $program = "process_name";
my $continue = 1;
$SIG{TERM} = sub { $continue = 0 };
while ($continue) {
my $status = `/bin/ps | /bin/grep $program | /bin/grep -v grep`;
if ($status) {
sleep 30;
} else {
my $pid = fork;
die "could not fork" unless defined $pid;
if ($pid) {
#give process a chance to start
sleep 5;
next;
}
exec $program; #child process replaces self with program
}
}
--
Chas. Owens
wonkden.net
The most important skill a programmer can have is the ability to read.
--
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
http://learn.perl.org/
