Thanks for the repost. This explains the difference in timeouts between version
4 and 5. I'm currently using version 5 but took the class for version 4. I'm
still not sure why a timeout of 1 second is being used when I configured Netview
for 2 seconds. Perhaps it learns that the device normally responds in under a
second and adjusts its value accordingly. Does anyone know how many outstanding
pings Netview can have on version 5? For version 4, I was told Netview could
only ping 10 addresses at a time. Is this configurable?
---------------------- Forwarded by Anthony D Philipp/Nationwide/NWIE on
2/17/1999 09:19:15 AM ---------------------------
Leslie Clark <lclark@us.ibm.com> on 02/16/99 11:46:14 PM
Please respond to Discussion of IBM NetView and POLYCENTER Manager on NetView
<NV-L@UCSBVM.ucsb.edu>
Sent by: Leslie Clark <lclark@us.ibm.com>
To: NV-L@UCSBVM.ucsb.edu
cc:
Subject: Re: Polling timeouts
Here's a re-post of a recent discussion of this that sounded good to me...
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
>I recently turned on netmon tracing in order to watch how netmon reacts to
a
>down interface. I have status polling defined for a 2 second timeout and
3
>retries. What I found in the trace log was three separate ping attempts
(I
>expected four). The first had a timeout of 1 second. The second ping had
a
>timeout of 2 seconds, and the third had a timeout of 1 second. Can anyone
>explain the timeout values observed, and why I did not see a thrid retry?
I
>have about 5000 objects being polled at 5 minute intervals by Netview 5.0
on AIX
>4.2.1.
Chris,
Here's an answer to the question about netmon's polling that you posted
December 10, 1998.
> I've noticed something interesting that I can't explain...
>
> On my Netview 5 system, I've got SNMP configured with one profile for the
> whole network. I'm using the following values:
>
> Timeout: 2.0
> Retry: 3
> Status: 5m
>
> And that's about it. According to my Netview class, this should result
in
> the following behavior:
>
> ping
> if no response, wait 2 seconds
> ping
> if no response, wait 4 seconds
> ping
> wait up to 8 seconds for response
> if no response then device is down.
>
That is the way it used to work in NetView v4
I'm not sure what formula was used in v5.0
Here's the new scoop for v5.1:
Given: Timeout=2, Retry=5
Initial Ping
Wait 2 seconds, ping
wait (2+1)=3 seconds, ping again
wait ((2+1)+1)=4, ping again
wait 2 seconds, ping last time
Total time (wasted) = 11 seconds
The scenario is:
Initial ping,
Wait time out value, ping
Wait time out value plus 1, ping
Keep waiting and adding 1, ping
Last timeout will be original timeout
Total number of pings sent will be the "Retry value"
Retry of zero will be interpreted as "one ping only" (wasn't that a line
from Red October).
=================================================
Don Davis - Consultant, Instructor
Alliance of Professionals & Consultants, Inc.
8600 Caswell Court
Raleigh, NC 27613-1100 USA Phone: 919-847-6056
Pager: 800-759-8888 PIN# 1462396
Email: Don@apc-services.com
Visit our web site at http://www.apc-services.com
=================================================
|