RE: [nv-l] Efficiency of snmpd.conf

["Barr, Scott" <Scott_Barr@csgsystems.com>]

You don't have much choice.

 

The smarset entries in ovsnmp.conf cannot be used during discovery. So if you DON'T use the communityNames.conf file, you will never get an SNMP request to work on them in the first place and thus you won't be able to have the smartset entry picked up in ovsnmp.conf. (Unless your smartset is based SOLEY on the name of the box. It's a chicken vs. egg problem. You have to have an entry in ovsnmp.conf first before the smarset criteria applies.

 

Someone chime in if this is wrong. Our smarset criteria is often based on SysContact and SysLocation which of course, can't be interrogated until after the box has been discovered which requires SNMP in the first place.

-----Original Message-----
From: Hill, Channing [mailto:CHill@BBandT.com]
Sent: Friday, July 25, 2003 12:00 PM
To: 'nv-l@lists.tivoli.com'
Subject: [nv-l] Efficiency of snmpd.conf

Netview 7.1.3 w/ fixpack 1 on AIX 5.1

I am in the process of building a new Netview system from scratch (complete overhaul from Netview 6.0.2). In the past, I've always had smart sets in my snmpd.conf for determining which community strings, polling settings, etc to use when the default community did not apply. These smart sets (2 of them) have about 2000 nodes a piece in them. I am thinking about letting the "communityNames.conf" file take care of this for my future build, but that would make my snmpd.conf file rather large with about 4000 entries in the "Specific Nodes" section. My question is: What would be more efficient for the interaction of netmon and the snmpd.conf file......to have these 4000 nodes in smart sets and then list the smart sets in snmpd.conf or have 4000 individual entries for each node in the snmpd.conf.

Channing Hill
EMS Analyst II
chill@bbandt.com
BB&T
Wilson, NC