Nagios Support Forum

Hello,

I am completely new to setting up nagios to monitor snmpd. I have a freebsd nagios box and centos machine running snmpd.

I used the exec to run check_mailq on the centos machine. I can use snmpwalk/snmpget to query it, but when I check_snmp found in nagios, it just gets plugin time out error. I'm sure its due to my lack of understanding or configuration problem. But I was wondering if someone could help me out please. I have googled the error some. But not getting headway in this way.

How can I use check_snmp to query the OID? (also, what is the $USER1 I see frequently referenced in naigos literature. Right now, I'm thinking it is the user nagios runs as maybe.)


check_snmp command:

(ran as root) check_snmp -H <ip address> -o <numeric_oid> -c <community string>
exec command in remote server snmpd.conf:
exec <name> <oid> <path/to/check_mailq + flags>

I also tried: check_snmp -H <ip address> -o <numeric_oid> -c <community string> + critical and warning flags and taking the flags off the exec statement


Please note, I can get the results with snmpwalk/snmpget. so I think its my understanding how to get Naigos to work with this. I aslo based the OID on the oids in the the example exec statements. Which seems to work.



Note: the error I get from the above is:

CRITICAL - Plugin timed out while executing system call

Can you try appending -t 60 or -t 120 t o the end of the check_snmp command? It could be simply timing out as the plugin runs. I think the default is only 10 seconds, so this would explain it. Especially since things are working with a regular walk.

Hello rkennedy,

Thanks for your help, I found that I can snmpwalk via DNS name but not IP address. This made me think that I should try it again with check_snmp with DNS name (I had been with IP address only). However, it still failed.

I tried adding -t 200, but still failed.

Here is my snmpwalk command: snmpwalk -v2c -c <community_string> <DNS hostname> <oid> which works,


but check_snmp doesn't work with (I've tried various flags, leaving some off, using it again:

check_snmp -H <DNS Hostname> -o <oid> -c <community_string> -P 2c

everything is always CRITICAL - Plugin timed out while executing system call

is there any debug logs?

Try adding the verbose flag to the command and see if you get better details of the failure when using the check_snmp plugin.

Hello tgriep,

Here are the results

# ./check_snmp -H <DNS host> -P 2c -o <OID> -v
/usr/local/bin/snmpget -Le -t 3 -r 5 -m '' -v 2c [authpriv] <DNS_HOST>:161 <OID>
CRITICAL - Plugin timed out while executing system call


to me, it appears that its timing out due to possible credential problems. But the odd thing is, why does snmpget/walk work directly (though this is calling snmpget) with the below

snmpget -v2c -c <community string> <dns host> <oid>

I've read that -m is not needed when using numeric OIDs

I've tried entering the line above
/usr/local/bin/snmpget -Le -t 3 -r 5 -m '' -v 2c [authpriv] <DNS_HOST>:161 <OID>

minus the port number and [authpriv]

I dont have enough exerpience to know how check_snmp is working different (yet)

OK,

I found the problem, it works now. I was using a lowercase 'c' vs capital 'C' for the community string.

Thanks for the help. (found via man page for check_snmp)


this might be there as well, but I need to find out how to translate this into the nagios interface

for the check command: check_snmp
$USER1$/check_snmp -H $HOSTADDRESS$ $ARG1$

I don't know if I just put the flags in ARG1? but that seems to generate an error when adding it to the host.

I found out how to define my own definition, now I can get it in a format like this: even figured out the error I was getting when restarting!

csptra wrote:I found out how to define my own definition, now I can get it in a format like this:

Code: Select all

$USER1$/check_snmp -H $HOSTADDRESS$ -o $ARG1$ -C $ARG2

even figured out the error I was getting when restarting!

You got it! The powerful thing about Nagios, is you can be as flexible as you want, or hardcode in all the variables as dynamic $ARGn$ variables.

Are we good to lock this thread up?

Hello,

I think so, I am working on writing my own wrapper to check_snmp so I can get a custom message and message count into nagios. (like check_qmailq)

other than that, it is just which 32 is the message count, but not very descriptive with SNMP critical.

You can use -l 'text here' to input text before the number 32, if that's what you're looking for.