What a good question! I was not thinking about that Smartset entry. In
effect you have a Smartset entry that includes a community that is not
correct for all members, and an overriding individual entry for just the
community. All of those bits of information are considered and stored on a
per-interface basis when netmon starts up, and re-evaluated when you make
changes to ovsnmp.conf. You can see it doing it if you turn on tracing.
So it should be possible for it to find the new record when it is looking
for a
community, but use the timeouts it has stored for the interface. You should
be
able to tell by watching the trace during a demandpoll. You will see it
timing
out and retrying. (It happens right at startup, so to see it you should
turn on
tracing using serversetup, rather than from the commandline.)
This case SHOULD be no different than one that does not involve the
community names file. What would happen if you had a specific entry
with, say, a 10 minute polling cycle, a Smartset entry with long timeouts,
and a default entry with a 5 minute polling cycle and short timeouts?
Would the node on the specific entry get a 10 minute polling cycle, and
long timeouts, which you want, or would it get a 10 minute polling cycle
and short timeouts? You should be able to see easily which case it is
by examining the netmon trace. If this scenario does not work as you
want it to, you might have a case to take to Support.
I hope you let us know what you find out.
Specifying communities on Smartset entries in ovsnmp.conf has always
been a little tricky, since so often membership in the Smartsets is based
on snmp criteria - a chicken/egg problem. I assume your Smartset is based
on something other than snmp attributes.
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
Detroit
"Peter Anderson" <pETERanderson@westpac.com.au>@tkg.com on 11/08/2000
10:51:19 PM
Please respond to IBM NetView Discussion <nv-l@tkg.com>
Sent by: owner-nv-l@tkg.com
To: "Netview List (E-mail)" <nv-l@tkg.com>
cc:
Subject: RE: [NV-L] Netview changing snmp config by itself
Hi Leslie,
Do I interpret what you are saying as the settings:
hostname1:xyz::::::::::::
will use the settings as defined for the smartset plus these?
Or is it that it uses the default settings plus these?
Regards,
Peter Anderson
Senior Communications Analyst
Ph: 0011 61 2 99025938
<Remove ETER from my address to reply>
Any views or opinions presented are solely those of the author and do
not necessarily represent those of Westpac Banking Corporation.
-----Original Message-----
From: SZEWCZYK, Jack
Sent: Thursday, November 09, 2000 2:31 PM
To: ANDERSON, Peter
Subject: FW: RE: [NV-L] Netview changing snmp config by itself
-----Original Message-----
From: Leslie Clark/Southfield/IBM
Sent: Thursday, 9 November 2000 3:24
To: IBM NetView Discussion
Cc:
Subject: RE: [NV-L] Netview changing snmp config by itself
-----Original Message-----
From: Leslie Clark/Southfield/IBM
Sent: Thursday, 9 November 2000 3:24
To: IBM NetView Discussion
Cc:
Subject: RE: [NV-L] Netview changing snmp config by itself
And it DID keep the other settings, right? It only overrode the
read community. Nothing else was filled in, as your example shows.
hostname1:xyz::::::::::::
It looks to me like the daily configuration poll probably
did what it was supposed to, and added entries for you.
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
Detroit
pETERanderson@westpac.com.au@tkg.com on 11/08/2000 08:59:45 PM
Please respond to IBM NetView Discussion <nv-l@tkg.com>
Sent by: owner-nv-l@tkg.com
To: "Netview List (E-mail)" <nv-l@tkg.com>
cc:
Subject: RE: [NV-L] Netview changing snmp config by itself
Further investigation found that yes, we did have names in
communityNames.conf.
There is conflicting information in the doco. Some doco says these names
are only used during the initial discovery, other doco says that there are
other situations:
1. After a SNMP timeout.
2. When a demand poll fails
Surely Netview would only attempt to use the alternates if:
1. The device was unreachable by SNMP but reachable by ping, OR
2. The device was unreachable by SNMP and it was configured to do SNMP
status polling.
Neither of those scenarios occurred here.
Also, once netmon had used an alternate it would only modify the snmp conf
if "public" failed and the alternate worked. Then and only then it should
modify the config by adding a new entry BUT IT SHOULD KEEP ANY OTHER
EXISTING SETTINGS. For example, timeouts or no. of routing entries.
Regards,
Peter Anderson
Senior Communications Analyst
Data Communications Support
Ph: 9902(5)5938
-----Original Message-----
From: SZEWCZYK, Jack
Sent: Thursday, November 09, 2000 9:41 AM
To: ANDERSON, Peter
Subject: FW: RE: [NV-L] Netview changing snmp config by itself
-----Original Message-----
From: Patel, Shaileshbhai B
Sent: Wednesday, 8 November 2000 22:37
To: 'IBM NetView Discussion'
Cc:
Subject: RE: [NV-L] Netview changing snmp config by itself
-----Original Message-----
From: Patel, Shaileshbhai B
Sent: Wednesday, 8 November 2000 22:37
To: 'IBM NetView Discussion'
Cc:
Subject: RE: [NV-L] Netview changing snmp config by itself
Leslie,
Looks like this is a bug. Here is the scenario. I have a network
device (Cisco Router)
which has a "public" read community set on it (This is a lab
environment). I setup
different community string in ovsnmp.conf using xnmsnmpconf
application. Just for an example say it was set to "notusepublic".
There is no community string
defined in /usr/OV/conf/communityNames.conf. That file is blank. If
I use demandpoll or configuration polling on that device, it is
modifying community string from "notusepublic" to public. I have
tested this for at least 5 times. If the community
string is set to something else then "public" on that device,
demandpoll/configuration polling fails. That is the way it is
supposed to work. My global default is set to "public". If demandpoll
defaults to Global default that is fine but it should not modified
entry in ovsnmp.conf file.
Any comments?
Shailesh Patel
EDS
-----Original Message-----
From: Leslie Clark/Southfield/IBM [mailto:lclark@us.ibm.com]
Sent: Friday, November 03, 2000 8:44 AM
To: IBM NetView Discussion
Subject: Re: [NV-L] Netview changing snmp config by itself
Spooky, eh? Well, there only 2 possible explanations. Either someone has
added a specific entry unbeknownst to you, or someone has added an entry to
/usr/OV/conf/communityNames.conf, which Netview will try when nothing in
ovsnmp.conf works. If something in there works, it adds entries to
ovsnmp.conf.
The Sherlock Holmes approach to problem determination works for me:
When you eliminate the impossible, whatever remains, however improbable,
must be the truth.
Cordially,
Leslie A. Clark
IBM Global Services - Systems Mgmt & Networking
Detroit
"Peter Anderson" <pETERanderson@westpac.com.au>@tkg.com on 11/02/2000
08:20:01 PM
Please respond to IBM NetView Discussion <nv-l@tkg.com>
Sent by: owner-nv-l@tkg.com
To: "Netview List (E-mail)" <nv-l@tkg.com>
cc:
Subject: [NV-L] Netview changing snmp config by itself
Hi,
We have Netview Unix 6.0.1 on Solaris 2.7.
I've had a problem occur a couple of times where Netview in its infinite
wisdom creates new entries in the ovsnmp.conf_db database for individual
nodes, using Netview defaults but with community names not found anywhere
except in the managed device itself.
For example, we use public for all our read-only monitoring but use say
"xyz" as our read-write community name.
I setup the ovsnmp.conf_db with only static entries for our two Netview
boxes, and either the default or a smartset entry for every other managed
node.
Our default is:
*.*.*.*:public:*:100:2:7200:::604800:604800:86400:10:1:0:1
Our smartset entries are similar but with slightly different timeout
settings, depending on the set.
branch_swch:public:*:200:2:7200:::604800:2419200:86400:1:0:0:0
I've found two occasions where entries are created like:
hostname1:xyz::::::::::::
It's only happened twice, and it has happened when nobody has been using
Netview, just the daemons in the background doing polling. Entries haven't
been created for EVERY managed node - just some.
Unfortunately, the above settings cause high CPU usage on our smaller Cisco
routers, which I think is due to netmon wanting to get the entire routing
table and with a small timeout, retries occur.
I Netview designed like this and if so why?
Regards,
Peter Anderson
Senior Communications Analyst
Westpac Banking Corporation
Ph: 0011 61 2 99025938
<Remove ETER from my address to reply>
Any views or opinions presented are solely those of the author and do not
necessarily represent those of Westpac Banking Corporation.
_________________________________________________________________________
NV-L List information and Archives: http://www.tkg.com/nv-l
_________________________________________________________________________
NV-L List information and Archives: http://www.tkg.com/nv-l
_________________________________________________________________________
NV-L List information and Archives: http://www.tkg.com/nv-l
_________________________________________________________________________
NV-L List information and Archives: http://www.tkg.com/nv-l
_________________________________________________________________________
NV-L List information and Archives: http://www.tkg.com/nv-l
_________________________________________________________________________
NV-L List information and Archives: http://www.tkg.com/nv-l
|