Dual-port monitoring

Israel Brewster israel at frontierflying.com
Tue Mar 18 18:44:54 CET 2008


On Mar 17, 2008, at 1:30 PM, Cook, Garry wrote:

> Well, maybe you can use check_cluster as the host check? It sounds  
> logical to me, but I’ve never used that check so I’m not really sure.
> Also, check_icmp has the ability to PING multiple addresses. Perhaps  
> using this as your host check against both IPs would give you the  
> results you’re looking for…

The good: check_icmp appears to give the desired ok/critical behavior  
when used with multiple hosts. check_cluster also seems to work, and  
perhaps somewhat more efficiently, since  it is just using the results  
from the service checks, rather than re-checking. Of course, this may  
not be desired.

The bad: check_icmp (with default settings, at least) shows 60% packet  
loss to the device- even though a concurrently running ping shows no  
packet loss. Setting the max packet interval setting to something high  
(at least 3 seconds, although even that isn't high enough some times)  
seems to help, but slows down the check. Setting a lower number of  
pings (1 or 2) also works, which may be acceptable for a host check,  
as the default is just a single ping anyway.

The Ugly: Creating a second service for the second WAN port seems to  
ignore the "no re-notifications" setting, thereby flooding my inbox  
with critical messages every so often (variable timing, usually about  
every half hour). I started a separate thread on that issue though.

Thanks for the suggestions. It looks like check_icmp as the host check  
is going to be the best one of the lot. Once I get it working  
properly :)

-----------------------------------------------
Israel Brewster
Computer Support Technician
Frontier Flying Service Inc.
5245 Airport Industrial Rd
Fairbanks, AK 99709
(907) 450-7250 x293
-----------------------------------------------
>
> Thanks,
> Garry
>
> From: Israel Brewster [mailto:israel at frontierflying.com]
> Sent: Monday, March 17, 2008 2:37 PM
> To: Cook, Garry
> Cc: Nagios Users Mailinglist
> Subject: Re: [Nagios-users] Dual-port monitoring
>
> On Mar 17, 2008, at 12:06 PM, Cook, Garry wrote:
>
>
> Yes, check_ifstatus is a service associated with the hosts. It  
> checks interfaces only. The host check is a check_icmp against the  
> host address, which is the loopback. If any of the interfaces are  
> down, check_ifstatus will report this, but the host does not show as  
> down as long as the loopback responds to a PING, which it will if  
> any of the interfaces are up.
>
> Ok, I see what you are saying, and it does sound as though it would  
> work as I desire. Unfortunately, for the device in question, it  
> doesn't look as though this will be an option, as I don't think I  
> can use loopback addresses (primarily Linksys RV082 devices). Thanks  
> anyway.
>
> -----------------------------------------------
> Israel Brewster
> Computer Support Technician
> Frontier Flying Service Inc.
> 5245 Airport Industrial Rd
> Fairbanks, AK 99709
> (907) 450-7250 x293
> -----------------------------------------------
>
> Thanks,
> Garry
>
> From: Israel Brewster [mailto:israel at frontierflying.com]
> Sent: Monday, March 17, 2008 1:46 PM
> To: Cook, Garry
> Cc: Nagios Users Mailinglist
> Subject: Re: [Nagios-users] Dual-port monitoring
>
> On Mar 17, 2008, at 11:03 AM, Cook, Garry wrote:
>
>
>
> Do you have the ability to use loopback addresses on these devices?  
> If so, that would be used as the host address, and then you could  
> use other checks for the interfaces. For instance, all of my Cisco  
> routers have a loopback setup. I then use ‘check_ifstatus’ to check  
> the individual interfaces.
>
> Thanks,
> Garry
>
>
> I assume the check_ifstatus for the interfaces are services  
> associated with the host? If so, then what do you use for your host  
> check? How do you get the host to show as down if all the interfaces  
> are down, but not if one or both (or more, if you have more than  
> two, although that is not the case with any of our devices) are up?
>
> -----------------------------------------------
> Israel Brewster
> Computer Support Technician
> Frontier Flying Service Inc.
> 5245 Airport Industrial Rd
> Fairbanks, AK 99709
> (907) 450-7250 x293
> -----------------------------------------------
>
> From: nagios-users-bounces at lists.sourceforge.net [mailto:nagios-users-bounces at lists.sourceforge.net 
> ] On Behalf Of Israel Brewster
> Sent: Monday, March 17, 2008 12:41 PM
> To: Nagios Users Mailinglist
> Subject: Re: [Nagios-users] Dual-port monitoring
>
> On Mar 17, 2008, at 9:03 AM, Gary Every wrote:
>
>
>
>
> In your services file:
>
> define service {
>   use                   generic-service
>   name                  Ping
>   host_name        multi_homed_server_name
>   service_description   Ping Second IFace
>   check_command         check_ping2!10.1.1.100!3000,10%!10000,20%
>   contact_groups        Unix
> }
>
> for your checkcommands:
>
> define command{
>         command_name    check_ping2
>         command_line    /usr/lib/nagios/plugins/check_ping -H $ARG1$  
> -w $ARG2$ -c $ARG3$ -p 5
>         }
>
>
> check_ping2 adds the first ARG as the ip address that you want  
> monitored - This means that the HOSTNAME will remain the same, but a  
> diff interface will be checked.
>
>
>
>
> Thanks for the response. If I understand your reply correctly,  
> though, this just adds a second service for the host, with a  
> separate IP from that specified in the host directive. it does not,  
> however, allow me to monitor the HOST on two separate IP's, taking  
> into account that if either interface is up, the host is up. Perhaps  
> my question wasn't quite clear enough. Take, for example, the  
> following situation:
>
> host a has two network interfaces: a primary with an IP of  
> 10.1.1.100 and a secondary with an IP of 10.1.1.101.
> Host a can be reached through either interface.
> Host a has its host address defined as the primary of 10.1.1.100,  
> with a check command of check_ping
> Host a also has a service defined as above, checking the secondary  
> interface (10.1.1.101)
>
> If the secondary interface goes down, then everything is fine:  
> nagios notices that the service for the secondary interface is  
> critical, checks the host on the primary, sees that as being fine,  
> and sends an alert about the service (secondary port), while leaving  
> the host in an OK state. This is exactly as it should be.
>
> However, now imagine that the primary interface goes down instead.  
> Nagios attempts to check the host on the primary interface, which  
> also fails (sicne the interface is down), and therefore nagios  
> alerts that the host is down, and any devices behind it are  
> unreachable. This, however, is incorrect-since the secondary  
> interface is still connected and fully functional, the host and  
> everything behind it is still up and reachable. What I want is a  
> situation where Nagios will send me an alert if either interface of  
> a host is down, but only consider the host to be down if BOTH  
> interfaces on the host are down. As I explained in my original  
> posting, I have considered a couple of options that may accomplish  
> this, but neither seems as elegant as I would like. Thanks for any  
> thoughts!
>
> -----------------------------------------------
> Israel Brewster
> Computer Support Technician
> Frontier Flying Service Inc.
> 5245 Airport Industrial Rd
> Fairbanks, AK 99709
> (907) 450-7250 x293
> -----------------------------------------------
>
> On Mon, Mar 17, 2008 at 9:38 AM, Israel Brewster <israel at frontierflying.com 
> > wrote:
> I have a number of devices on my network that have dual-interfaces
> such that either one can go down, but the device itself, and all the
> devices behind it (in the case of a router) is still up and reachable.
> What is the best way to set up monitoring of such devices, considering
> that as far as I can tell you can only assign one IP per host? I have
> looked at check_cluster, but if that is the best method I am unsure of
> the best way of implementing it. Things I have thought of:
>
> 1) Create two "dummy" hosts which are the two ports, and then a third
> host with a check command of check_cluster that looks at those two
> dummy hosts. This seems overly complicated, not to mention cluttering
> up my configs and nagios interface with three hosts where there is
> really only one.
>
> 2) Simply monitor the two ports as two separate hosts. This could
> work, but again clutters things up with multiple hosts where only one
> exists, and adds the requirement of multi-parenting any child devices,
> which can get ugly, especially in a graphical representation of the
> network.
>
> I have read http://nagios.sourceforge.net/docs/2_0/clusters.html, but
> that doesn't seem to apply, at least not directly, in that it refers
> either to situations where you have redundant services on one or more
> hosts (perhaps that is sort of what I want?) or to where you have
> multiple redundant hosts, but not to the situation where you have one
> host offering one service over two ports. The main problem with that
> doc, though, is that the cluster is always a service, which wouldn't
> work in my situation, as the host itself is a cluster. Any thoughts?
> Thanks.
> -----------------------------------------------
> Israel Brewster
> Computer Support Technician
> Frontier Flying Service Inc.
> 5245 Airport Industrial Rd
> Fairbanks, AK 99709
> (907) 450-7250 x293
> -----------------------------------------------
>
>
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by: Microsoft
> Defy all challenges. Microsoft(R) Visual Studio 2008.
> http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
> _______________________________________________
> Nagios-users mailing list
> Nagios-users at lists.sourceforge.net
> https://lists.sourceforge.net/lists/listinfo/nagios-users
> ::: Please include Nagios version, plugin version (-v) and OS when  
> reporting any issue.
> ::: Messages without supporting info will risk being sent to /dev/null
>
>
>
> -- 
> Gary Every
> "Pay it Forward!"
>
>
> NOTICE: This e-mail and any files transmitted with it are the  
> property of ARCADIS U.S., Inc. and its affiliates. All rights,  
> including without limitation copyright, are reserved. The  
> proprietary information contained in this e-mail message, and any  
> files transmitted with it, is intended for the use of the  
> recipient(s) named above. If the reader of this e-mail is not the  
> intended recipient, you are hereby notified that you have received  
> this e-mail in error and that any review, distribution or copying of  
> this e-mail or any files transmitted with it is strictly prohibited.  
> If you have received this e-mail in error, please notify the sender  
> immediately and delete the original message and any files  
> transmitted. The unauthorized use of this e-mail or any files  
> transmitted with it is prohibited and disclaimed by ARCADIS U.S.,  
> Inc. and its affiliates.
>
>

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.monitoring-lists.org/archive/users/attachments/20080318/ee264db1/attachment.html>
-------------- next part --------------
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
-------------- next part --------------
_______________________________________________
Nagios-users mailing list
Nagios-users at lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/nagios-users
::: Please include Nagios version, plugin version (-v) and OS when reporting any issue. 
::: Messages without supporting info will risk being sent to /dev/null


More information about the Users mailing list