We block ICMP 'from' the router to the LAN that supports the network on
which the IM Service is running. At this time I don't log those events to
syslog. We see no ICMP traffic emanating from the LAN where the service
resides to the router (I do block with two exceptions and log that). The
Nagios device is not shielded from ICMP (though it could and will be as of
now). We were seeing some high traffic from two customer networks at about
the same time. We will investigate that.

Thanks for the suggestion Mel.

- Mike

-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Mel Beckman
Sent: Tuesday, November 04, 2003 10:04 AM
To: InterMapper Discussion
Subject: RE: Nagios false down reads

Mike,

   I've been seeing problems like this in networks that turns out to be the
Blaster and Nachi virii in their "broadcast" mode -- when they try to ping
every IP address on the local subnet, which results in massive broadcast
packets on large (e.g. 10.0.0.0/8) subnets. Many routers discard ICMP
packets when faced with so much broadcast traffic. I don't believe
Intermapper will sense the broadcast traffic, because it may not be a high
bits per second data rate, just a high packets per second.

Have you put a sniffer on your network to see if you've got a broadcast
flood going on?

  -mel

>I rarely am watching the screen at the exact moment when the probe is 
>reported as down, so I can't tell you the "Reason" from the Satus Window.
>Here's what I have right now.
>
>The Debug log has nothing for the probe.
>
>The outage log looks like this:
>
>Mon, Nov 03, 2003      03:58 PM        Cheyenne Temp   29 seconds      Mon,
>Nov 03, 2003   03:58 PM
>Mon, Nov 03, 2003      04:01 PM        Cheyenne Temp   28 seconds      Mon,
>Nov 03, 2003   04:01 PM
>Mon, Nov 03, 2003      05:30 PM        Cheyenne Temp   28 seconds      Mon,
>Nov 03, 2003   05:30 PM
>Mon, Nov 03, 2003      07:20 PM        Cheyenne Temp   58 seconds      Mon,
>Nov 03, 2003   07:19 PM
>Mon, Nov 03, 2003      07:42 PM        Cheyenne Temp   28 seconds      Mon,
>Nov 03, 2003   07:42 PM
>Mon, Nov 03, 2003      07:59 PM        Cheyenne Temp   28 seconds      Mon,
>Nov 03, 2003   07:59 PM
>Mon, Nov 03, 2003      09:45 PM        Cheyenne Temp   28 seconds      Mon,
>Nov 03, 2003   09:44 PM
>Mon, Nov 03, 2003      11:13 PM        Cheyenne Temp   28 seconds      Mon,
>Nov 03, 2003   11:13 PM
>Tue, Nov 04, 2003      12:00 AM        Cheyenne Temp   28 seconds      Mon,
>Nov 03, 2003   11:59 PM
>
>  The event log looks like this:
>
>11/03 19:19:33  DOWN  Cheyenne Temp::  (Was up for 1 hour, 49 minutes, 
>2
>seconds)
>11/03 19:20:31  UP    Cheyenne Temp::  (Was down for 58 seconds)
>11/03 19:42:03  DOWN  Cheyenne Temp::  (Was up for 21 minutes, 32 seconds)
>11/03 19:42:31  UP    Cheyenne Temp::  (Was down for 28 seconds)
>11/03 19:59:03  DOWN  Cheyenne Temp::  (Was up for 16 minutes, 32 seconds)
>11/03 19:59:31  UP    Cheyenne Temp::  (Was down for 28 seconds)
>11/03 21:44:32  DOWN  Cheyenne Temp::  (Was up for 1 hour, 45 minutes, 
>1
>second)
>11/03 21:45:00  UP    Cheyenne Temp::  (Was down for 28 seconds)
>11/03 23:13:02  DOWN  Cheyenne Temp::  (Was up for 1 hour, 28 minutes, 
>2
>seconds)
>11/03 23:13:30  UP    Cheyenne Temp::  (Was down for 28 seconds)
>11/03 23:59:32  DOWN  Cheyenne Temp::  (Was up for 46 minutes, 2 
>seconds)
>
>-----Original Message-----
>From: [EMAIL PROTECTED]
>[mailto:[EMAIL PROTECTED] On Behalf Of Christopher L.
>Sweeney
>Sent: Tuesday, November 04, 2003 8:53 AM
>To: InterMapper Discussion
>Subject: Re: Nagios false down reads
>
>At 8:30 AM -0700 11/4/03, Mike Lieberman wrote:
>>Are packets from Nagios devices easily lost, or occasionally not sent,
>  >for some reason?
>  >
>  >We are using a remote temp probe which supports Nagios via IM. The  
> >Nagios code seems to work, and the temps are properly read. There is  
> >good open bandwidth between the probe and the IM service, (by a DS3
>>which during these down events may have under 10% utilization). Never 
>>the less we see frequent "down" events. Setting the timeout to a 
>>longer
>setting just makes the "Down"
>>longer before the "UP". It doesn't appear to be a delayed packet as 
>>much as a missed, unsent or discarded packet.
>>
>>Does anyone have anything else similar happening with Nagios supported 
>>equipment under IM?
>>
>
>What does the status window give as a reason for these down events?
>It would be worth checking the debug log, too, to see if IM is having 
>difficulty setting up and executing the Nagios command from time to time.
>
>-- Christopher
>
>--
>================================================
>Christopher L. Sweeney
>[EMAIL PROTECTED]
>http://www.dartware.com/
>
>____________________________________________________________________
>List archives:
>http://www.mail-archive.com/intermapper-talk%40list.dartware.com/
>To unsubscribe: send email to: [EMAIL PROTECTED]
>
>
>____________________________________________________________________
>List archives:
>http://www.mail-archive.com/intermapper-talk%40list.dartware.com/
>To unsubscribe: send email to: [EMAIL PROTECTED]


____________________________________________________________________
List archives: 
http://www.mail-archive.com/intermapper-talk%40list.dartware.com/
To unsubscribe: send email to: [EMAIL PROTECTED]


____________________________________________________________________
List archives: 
http://www.mail-archive.com/intermapper-talk%40list.dartware.com/
To unsubscribe: send email to: [EMAIL PROTECTED]

Reply via email to