I put $ against the node ip address in the seed file and then I changed
community string in NetView (ovsnmp.conf). I got node down as well as
interface down event for that node. Then I changed back the community
string. In the next polling cycle I got node up as well as interface up
event but the color of the icon still remains white (unreachable). Any idea
why? I waited couple of polling cycle to make sure that it does not become
green by itself. I did the manual ping from netview and icon became green
again. So the question is shouldn't it become green automatically again once
staus by SNMP is OK again?
I will still try again by modifying oid_to_type file. I will discover new
nodes after I modify oid_to_type file.
Thanks for all of your responses.
Shailesh Patel
Enterprise Management Services
Electronic Data Systems
Troy MI 48098
(248) 265-7573
-----Original Message-----
From: lclark@us.ibm.com [mailto:lclark@us.ibm.com]
Sent: Friday, August 04, 2000 12:39 AM
To: IBM NetView Discussion
Subject: RE: [NV-L] Status By SNMP
I would expect that things set in the oid_to_type file would be respected
on newly discovered nodes, or at the very best, after an ovtopofix.
It does sound as if netmon is not yet using the new approach. You can
tell if you do a demandpoll. The output shows the ifOper an ifAdmin
status.
Idea:
This week's customer is using the new feature in an interesting way. They
are doing ALL polling of servers with snmp (by the seedfile method) so
they can get a greater indication of the status of the OS than the ping
gives.
They are doing this on a 1-minute polling cycle. It is performing very well
since it has plenty of resources and a high-speed network connection. Note
that we have been very careful to unmanage all interfaces that we do not
really care about withing the group being polled in this way.
Warning:
I understood it to be the case that the $ in the seedfile did not change
discovery, only the polling method; however, I believe I saw it discover
everything when I added $*.*.*.* so went back to an explicit restrictive
list
of addresses, followed by the same list with $ in front of the addresses.
I'll have to investigate this more closely when I have a chance. Anyone
else seen this?
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
Detroit
Mark_LaFever@CARGILL.COM@tkg.com on 08/03/2000 05:32:07 PM
Please respond to IBM NetView Discussion <nv-l@tkg.com>
Sent by: owner-nv-l@tkg.com
To: nv-l@tkg.com
cc:
Subject: RE: [NV-L] Status By SNMP
If you changed the community strings on the devices and not in NetView,
you should get authentication failures from the devices when NetView
polls them. The authentication failure traps may show what device is
polling them with the wrong community string(should be your NetView
server).If the traps do not show the offending device you still should
be able to tell if netmon is doing the polling by the interval that you
receive the traps(They will be at the same interval that netmon is set
to poll).
Hope this helps,
Mark LaFever
Network Analyst
Excel Corp.
(316)291-2926
-----Original Message-----
From: shaileshbhai.patel@eds.com [mailto:shaileshbhai.patel@eds.com]
Sent: Thursday, August 03, 2000 1:37 PM
To: nv-l@tkg.com
Subject: FW: [NV-L] Status By SNMP
Hello all,
As per NetView version 6.0 release notes netmon can poll status by
sending SNMP queries for the ifAdminStatus and ifOperStatus values of
interfaces. One of the three ways you can do this is
By using the P switch in the /usr/OV/conf/oid_to_type file against
classes of devices.
I have modified oid_to_type file. How can I find out that netmon
is
polling these devices by SNMP and not by ping?
I tried changing the read community string on those devices but it
does not give node down or any kind of status down events.
Any idea?
Thanks in advance
Shailesh Patel
Enterprise Management Services
Electronic Data Systems
Troy MI 48098
(248) 265-7573
________________________________________________________________________
_
NV-L List information and Archives: http://www.tkg.com/nv-l
|