Interfaces stay in unknown state after nodes return to "up" state

Question

Hello

After looking through similar posts, I wasn't able to find a solution for my particular problem.

We monitor a large number of Cisco devices located in remote locations across country. At least several times a day many of them have short power outages/connectivity losses and so on. Nodes go down, a map displays that to us, everything is ok. Then as a node comes back up and restores connectivity, Orion displays its interfaces (and cpu and other snmp-based values) as "unknown". It only reports the state as back up, based on ping connectivity. Absolutely all our nodes have this problem.

It is easily resolved by restarting the "SolarWinds Network Performance Monitor" service in Service Manager (only this service is enough). But doing that several times a day, as often as our nodes go down, is not a good solution. Event viewer in Windows does not display any particular errors, like in this post -   - so that seems to be not our problem.

Also, the issue started about three weeks ago, before that NPM was working fine without this issue.

Version is 9.1.0

Thanks for any help!

FormerMember · Answer

When I selected yes v1, the process continued, my interfaces and the universal pollers were there.

Radioman, so that means that changing snmp version to v1 helped resolve the issue of interfaces going into "unknown" state?

Radioman · Answer

I am having the same problem  on a Solaris Server (essentially Unix/LINUX) operating a Broadsoft App. SEE

Previously set up as v2c after a release change, reboot or similar event, the universal mibs continue to collect data but the interfaces all go UNKNOWN.

As a test, I deleted the node and re-added it and it refused to recognized v2c, even though that's how it was installed initially and the Mibs didn't seem to have an issue. When adding a node the first step is of course to list resources (interfaces) and that's the rub, for some reason the interfaces no longer liked v2c and the nag popped up stating v2c UNREGCOGNIZED do you wish to switch to v1 yes/no.

When I selected yes v1, the process continued, my interfaces and the universal pollers were there. Since the univesals are both v1 and v2c I approached the lab and requested to find out if the snmpd.conf file had v2c open. This is what they sent back. I'm concerned that the SLASH / between v1 and v2c stops Solarwinds from from seeing the statement correctly. Need a SUPER Geek help on this one Solarwinds or Forum members with a solution, I can't go into production and delete and re-add 40 more servers I think I'm asking for trouble to even attempt that method.

############################################################################# snmpd.conf##   - created by the snmpconf configuration program############################################################################# SECTION: Access Control Setup##   This section defines who is allowed to talk to your running#   snmp agent. # rocommunity: a SNMPv1/SNMPv2c read-only access community name#   arguments:  community [default|hostname|network/bits] [oid]

FormerMember · Answer

All nodes run SNMP v2c. All nodes are affected.