SNMP errors on Solaris v10 monitored host after 10.2 upgrade

After upgrading from NPM 10.1.3 to 10.2 we began to see every 30 minutes a stream of error messages on our Solaris 10 host, repeating approximately 36 times:

Nov 30 20:51:39 tlc-host /usr/lib/snmp/snmpdx: [ID 702911 daemon.error] community_check() : bad community from localhost

Nov 30 20:51:39 tlc-host /usr/lib/snmp/snmpdx: [ID 702911 daemon.error] session_open() failed for a pdu received from localhost.36257

...

The 30 minute timeframe matches the rediscovery interval for the node., and had no issues prior to the Orion software upgrade.

Taking a Wireshark trace shows that these errors are triggered by a GetBulkRequest to OID 1.3.6.1.2.1.55.1.8.1.1. According to the SolarWinds Engineer Toolset MIB Browser this is the MIB-2 tree for ipv6MIB (55); we do not run IPv6, and no SNMP community name issues for any other SNMP requests from Orion. Using the MIB Browser I can reproduce the errors by using the Get Tree on this MIB.

Is there any way of elminating these errors from showing on our Solaris operator's console? The computer system operators are freaking out.

Find more posts tagged with

souchakov@infohedge.net

npm_10.2

solaris_10

Accepted answers

zizi

Orion started polling all IP addresses assigned to managed nodes. Addresses are stored in NodeIPAddresses table in Orion database. This is a new feature of Orion NPM 10.2. Orion is able to utilize DNS and SNMP or WMI to get this information.

For all SNMP Nodes:

Orion polls ipAdEntAddr (1.3.6.1.2.1.4.20.1.1) OID to get list of IPv4 addresses

Orion polls ipv6AddrAddress (1.3.6.1.2.1.55.1.8.1.1) OID to get list of IPv6 addresses

My understanding of the issue is, that Solaris SNMP agent is sending syslog messages when a SNMP manager asks for OID, which is missing in it's MIB. I'm not familiar with SNMP configuration on Solaris systems, but I would say it should be possible to suppress this syslog message.

All comments

GZhytar

HI telecom116,

please capture Wireshark trace while reproducing the problem and open support ticket so Dev team can investigate this one

thanks

danielleh

Telecom--

Would you please post back with your ticket # as well?

Thank you,
DH

FormerMember

A support case #292606 has been opened, and have sent Wireshark capture printouts for review.

infohedge

we are seeing the same problem after upgrade to 10.2,

do you have any update?

Dec 6 15:13:27 bsrv1101 /usr/lib/snmp/snmpdx: [ID 702911 daemon.error] community_check() : bad community from localhost

Dec 6 15:13:27 bsrv1101 /usr/lib/snmp/snmpdx: [ID 702911 daemon.error] session_open() failed for a pdu received from localhost.63077

Our snmpd.conf:

# cat /etc/snmp/conf/snmpd.conf

read-community $comsnmp

FormerMember

Glad (sort-of) that someone else is seeing this issue. Of course I don't have any pcap files of what it looked like before the upgrade to 10.2.

No resolution at this time, but I only opened a formal ticket around 11 am CST.

Besides sending my Wireshark SNMP polling info, a few hours ago the agent requested diagnoistics file, so someone is definately looking at it.

Will try to update as I learn more.

If you're under maintenance, you might want to open a ticket, too; the more the merrier. Also, it would have another customer pushing the issue.

danielleh

Hi infohedge--

Would you also please open a support ticket?

Thanks,
DH