Nagios check_http checking private IP

[srikanthm]

Hi I set up two hosts (host1, host2) and check_http service for each and when I test them:

Code: Select all

# ./check_http -H host1publicIP -p 9080 -w 5 -c 10
HTTP OK: HTTP/1.1 200 OK - 7909 bytes in 0.002 second response time |time=0.001918s;5.000000;10.000000;0.000000 size=7909B;;;0

When I added the new host - host2, I copied the above service and changed the ip and port to the new service. When I test it - works fine as below:

Code: Select all

# ./check_http -H host2publicIP -p 8080 -w 5 -c 10
HTTP OK: HTTP/1.1 200 OK - 7909 bytes in 0.003 second response time |time=0.002571s;5.000000;10.000000;0.000000 size=7909B;;;0

But when the command is executed as part of the nagios service - I get an error for host2 service:

Code: Select all

connect to address host2privateIP and port 8080: Connection refused
HTTP CRITICAL - Unable to open TCP socket

I do not have any problem with host1: check_http or other services of host2 - only check_http fails.

Here are the complete service descriptions:

host 1:

Code: Select all

define service{
        use                     generic-service
        host_name               host1
        service_description     DEMO-APP
        check_command           check_http!-H host1publicIP -p 9080 -u http://host1publicIP:9080/login/auth -w 5 -c 10
        max_check_attempts      1
        contact_groups          Escalation
        }

host 2:

Code: Select all

define service{
        use                     generic-service
        host_name               host2
        service_description     APP1
        check_command           check_http!-H host2publicIP -p 8080 -u http://host2publicIP:8080/login/auth -w 5 -c 10
        contact_groups          Escalation
        max_check_attempts      1
        }

Based on the error, nagios is trying to connect using private IP rather than public IP provided??
To verify this, I tested using the private IP and got the same error at prompt:

Code: Select all

# ./check_http -H host2privateIP -p 8080 -w 5 -c 10
connect to address host2privateIP and port 8080: Connection refused
HTTP CRITICAL - Unable to open TCP socket

Anybody can advise how to fix this??

thanks in advance.

[jdalrymple]

You'll either need to modify your command definition for check_http to use your arguments or modify your host definition to have the public IP of the machine. The check by default uses the macro $HOSTADDRESS$ which presumably is your private IP.

As an aside, the -H argument isn't used in the check_http plugin.

[srikanthm]

jdalrymple, thanks for your quick response.

I tried playing with the command in different ways and solved the issue. Instead of "-H" I used "-I" and it worked.

But still the big question why it works for one host but not other is not answered.

Also -H argument is available according to usage help, atleast in the version I am using:

Code: Select all

Usage:
 check_http[b] -H <vhost>[/b] | -I <IP-address> [-u <uri>] [-p <port>]
       [-J <client certificate file>] [-K <private key>]
       [-w <warn time>] [-c <critical time>] [-t <timeout>] [-L] [-E] [-a auth]
       [-b proxy_auth] [-f <ok|warning|critcal|follow|sticky|stickyport>]
       [-e <expect>] [-d string] [-s string] [-l] [-r <regex> | -R <case-insensitive regex>]
       [-P string] [-m <min_pg_size>:<max_pg_size>] [-4|-6] [-N] [-M <age>]
       [-A string] [-k string] [-S <version>] [--sni] [-C <warn_age>[,<crit_age>]]
       [-T <content-type>] [-j method]

[jdalrymple]

Sorry, I was unclear: -H i s available, it's just not really used.

Code: Select all

define command {
       command_name                             check_http
       command_line                             $USER1$/check_http -I $HOSTADDRESS$ $ARG1$
}

I can't be certain that this is what your command looks like, but it is likely. This is what Nagios actually *does* to check your http server.
$HOSTADDRESS$ expands to the IP of the host you're checking
$ARG1$ expands as everything between the 1st and 2nd "!" symbol in your check_command

Based upon the initial service definitions you provided...
For host 1 the actual command being ran to check your host is:

Code: Select all

/somepath/check_http -I <host1-some-IP> -H <host1-pub-IP> -p 9080 -u http://host1publicIP:9080/login/auth -w 5 -c 10

For host 2 the actual command being ran to check your host is:

Code: Select all

/somepath/check_http -I <host2-some-IP> -H <host2-pub-IP> -p 8080 -u http://host2publicIP:8080/login/auth -w 5 -c 10

Perhaps you would get more predictable results based upon your command definitions if you removed the "-I $HOSTADDRESS$" portion of the command definition.