Our environment is SolarWinds Orion Core Services 2011.2.2 polling 2252 elements. Currently it is polling at 21% of its maximum rate.
I have noticed on a daily basis that several of our physical server computers being monitored will show up in the Nodes With Problems section. It states either "One or more Interfaces are in an Unknown state" or "Interface '<interface name>' is in an Unknown state." When this is observed, nothing appears to be wrong with the servers in question and no related errors are logged. If I restart SNMP (an any dependent services) on the listed servers, these computer disappear from the Nodes With Problems list within a few minutes. However, this is only a temporary condition. Most of the computers will return to the list the following day, while the rest will return in less than an hour.
I have noticed that all of the computers that have this problem are running HP ProLiant server hardware with a Windows 2003 operating system. No HP ProLiants running Windows 2008 or 2008 R2 do this in our environment. We don't have any Dell's running 2003, but one of the problem servers used to be on IBM hardware until it died and was cloned onto HP hardware and now this problem appears when it did not previously. That said, not all of our HP ProLiants running 2003 experience this issue, but most do.
The servers are generally running HP ProLiant Support Pack (PSP) v8.70, though some are running older versions. In the case of the one server that used to be an IBM, it does not have any HP software installed.
Does anyone know what the cause of this problem is or how to stop it from occurring (save for scripting SNMP to restart daily)?
Does anyone have a solution to this issue? It seems relatively common.