nv-l
[Top] [All Lists]

Antw: Ruleset using RESOLVE and FORWARD

To: nv-l@lists.tivoli.com
Subject: Antw: Ruleset using RESOLVE and FORWARD
From: "Michael Seibold" <Michael.Seibold@gek.de>
Date: Fri, 01 Sep 2000 09:41:09 +0200
Hi Ray,

I use another approach for this issue: if you can wait fro three minutes before 
getting the IF_DOWN - event then you can modify the polling paramters so that 
"down" will only occure after netmon was pinging for three minutes without 
success. Go to "Options -> SNMP Configuration" and change the "Timeout"-Value 
and "Retry Count". Remember that the timeout-period is doubled for every retry:

by example: Timeout 2, number of retry's sets total time until IF_DOWN - event
 
initial ping       wait 2 seconds         total time 2 seconds till IF_DOWN - 
event
retry  1            wait 4 seconds         total time 6 seconds  till IF_DOWN - 
event
retry  2            wait 8 seconds         total time 14 seconds till IF_DOWN - 
event
retry  3            wait 16 seconds       total time 30 seconds till IF_DOWN - 
event
retry  4            wait 32 seconds       total time 62 seconds till IF_DOWN - 
event
retry  5            wait 64 seconds       total time 126 seconds till IF_DOWN - 
event
retry  6            wait 128 seconds     total time 254 seconds till IF_DOWN - 
event

Let's look at the probable reasons for those fake downs. One possibility is 
that the timeout/retry combination specified in netview is not appropriate.
If the log of those devices shows interface down/up for short times, then there 
is a real problem and it's better to solve this rather to tell netview to 
ignore it.
If you look at the log of the devices with fake downs and they show no 
interface down, then they were too busy to answer the netmon ping request or 
there was a network problem between your netview and the device.

My experience is that with timeout 1.8 and retry 5 in a local network and a 
frame-relay network with a maximum of 3 hops those reoccuring "fake" interface 
downs are showing those devices which will become critical within the next few 
weeks because they are permanently overloaded (speaking of cisco devices). You 
may not see this looking at the CPU-load of these devices or the load on single 
interfaces, but that's my experience. Another cause for those "fake" interface 
downs are routing-problems, by example updates of large routing tables 
reoccuring very fast due to a flippy interface. So I would play with the 
timeout and retry parameters a bit and then look for the reasons if those fake 
downs still exist.


Hope this helps

Michael Seibold



>>> Ray.Foss@motorola.com 01.09. 2.03 Uhr >>>
Forgive me if this is a FAQ, my ears are still wet when it comes to NV6K.

I'm running NetView 6.0 and forwarding OV_IF_Down and OV_IF_Up events to a
3.6.2 TEC based on a NetView ruleset.  My question is:

Can I use the RESOLVE template, as in the sample (sampcorrIuId.rs) ruleset,
to suppress quick down/up indications that are not real outages?  Here is
the pseudo-code for my desired results:

if you get a OV_IF_Down
        Hold it for a while (3 minutes)

if you get a OV_IF_Up
        if you have a OV_IF_Down in hold from the same IP
                resolve this event

if the Hold timeout expires
        forward the OV_IF_Down event

I also don't know if I can cascade rule sets.   Any help if greatly
appreciated.  Thanks.

-- 
~~~~~~~~
Ray Foss
_________________________________________
mailto:Ray.Foss@motorola.com 
Office: 480-441-1093 Mobile: 602-721-4792
 Pager: 800-759-8352 PIN: 1244994
   FAX: 480-441-5455

Motorola, Inc.
Global Computing & Telecommunications
8111 East McDowell Road - AZ33-H1780 
Scottsdale, Arizona 85257 USA
_________________________________________
_________________________________________________________________________


<Prev in Thread] Current Thread [Next in Thread>
  • Antw: Ruleset using RESOLVE and FORWARD, Michael Seibold <=

Archive operated by Skills 1st Ltd

See also: The NetView Web