[Nagiosplug-help] Problem with using check_ping sometimes; pipe problems

[Andreas Ericsson]

Sebastian Berm wrote:
> <snip>
> 
>>>I keep getting the 'Could not open pipe' errors at random intervals.
>>>
>>>Does anyone perhaps know why this happens ?
> 
> 
> [Andreas Ericsson] 
> 
>>Most likely it's because your system is starved of file-descriptors. You
>>could try either upping the ulimit -n value of Nagios or use check_icmp
>>instead. check_icmp uses a single socket rather than a pipe (two
>>file-descriptors fewer per check). It's also a lot faster, so resources
>>aren't hogged so long.
>>
> 
> 
> Thank you.
> I'm going to try to up the ulimit for the time being, and if it doesn't
> work, I'll take a look at migrating to check_icmp. It's quite a lot of
> hosts, and the system is going into production 'mode' about a week ahead of
> schedule (yeah, that too can happen with IT projects).
> 

Umm... The syntax is compatible (except that check_icmp doesn't support 
long options), so all you have to do is change your command object 
definition to read

define command{
    command_name  check_ping
    command_line  $USER1$/check_icmp -H $HOSTADDRESS$ -w $ARG1$ -c $ARG2$
}

If you want to make the switch on a running Nagios you can simply copy 
the check_icmp binary to $USER1$/check_ping. That way you can try them 
out in a manner of minutes, and restore the old check_ping if it doesn't 
work to your liking.

-- 
Andreas Ericsson                   andreas.ericsson at op5.se
OP5 AB                             www.op5.se
Tel: +46 8-230225                  Fax: +46 8-230231