This discussion has been locked. The information referenced herein may be inaccurate due to age, software updates, or external references.
You can no longer post new replies to this discussion. If you have a similar question you can start a new discussion in this forum.

Interface in "Unknown State"

I have a Dell PowerEdge 2950 (Tower) that is giving me a periodic "Unknown State" error. It doesn't appear to stop Solarwinds from receiving SNMP data, but it does consistently show up in "Nodes with Problems". None of my other server have this issue, and the only difference between this server and the others is that it is a tower versus a rack server.


Typically, I can restart the SNMP service on the server and it kicks back in, but it will return within a few days. The NIC model as read by Orion is: Broadcom BCM5708C NetXtreme II GigE (NDIS VBD Client). I've seen limited information around other forums referencing this problem with Broadcom nics, but I'm all out of ideas.


 Has anyone else been experiencing similar issues?

  • Well, I thought I'd bring this up again. I'm still at a loss - every so often these NICs (now, I've even seen an Intel NIC do it occasionally) will go into an "Unknown State" and I stop receiving CPU Load, Memory, and generally any SNMP information. I have to restart the SNMP service in order to receive data again. This happens on several of my Dell servers. Could it be related to the Dell SNMP Client software?

  • I have seen the same problem with a number of our Dell PE's (different models and ages), although not all, and some of them more frequent than others. I never to the time to correlate hardware/software to find a common factor.


    The only time we could reproduce it, was after installing the monthly MS patches that also leaves the SNMP service in an unresponsive state, unless you reboot the server (or restart the service).


    Little over a year ago, we decided to switch to IBM hardware, that do not give me the problem, although they also have the BroadCom NetXtreme NIC's on board.


    We also decided to part with Dell OpenManage in favor of IBM Director. This required us to leave to DOM client on the server, but also install the IBMD client. After that the occurrence of the problem seems to have all but disappeared. I do not want to draw any conclusions, but it may be that the IBMD client replaced a common dll.


    So your question whether it may be the Dell SNMP client, I would say that is possible that one of the installed components related to the DOM client may be 'less robust'.


    An obvious advise is to ensure that you installed the latest firmware and software updates from Dell.

  • I am experiencing the same problem as well I have 5 Supermicro servers plugged into a Dell 3348 switch. 3 of the servers have Intel NICs and 2 of them have Broadcom BCM57xx nics. I constently see this same behavior with the Broadcom NICs. has anyone tried updating the NIC driver?

  • I have! Since I posted this, all my Dell Servers have started to do this. Not as frequently on the Intel NICs, but still does occur. I've updated to the latest firmware and drivers, but nothing helps.


    I've gone as far as writing a script that will restart the SNMP service on the systems when they go into an unknown state, but I'm having difficulties making it run properly.


    I was thinking this problem is pointing more and more towards the Dell OpenManage SNMP agent, but now with your comments about the Broadcom NICs, I'm lost again.

  • Hi all,


     I have a few PowerEdge servers with Broadcom NICs as well.  What OID can be used to monitor the physical state of the adapters?


    For example if I unplug a network connection from the second port of the network adapter, it would report a link fail?


     Thanks,

  • I get the same "unknown status" interfaces on our Cisco 6500 series core switches.  Does not matter if its a VLAN or physical Ethernet interfaces but looks to happen on the same ones every day.  My temporary fix is to go to "Manage Nodes" --->Select the interfaces that in the "unknown state" & grayed out ---->click on the "More Actions" drop down menu and select "Poll Now".  this usually works for about a while but by the next day they'll be back to "unknown status" 

    My first thought is that I have to many devices/interfaces that I'm polling but if so, why the same ones every day?

    Thanks,

  • Hi,

    Have the same prob on our 6500 switches after upgrading to v10.0

    -Kurt

  • Hi again,

    Forgot to add: Interface Polling stopped. Could not remap interface. Check ifDescription change

    -Kurt

  • snmp ifmib ifindex persist on the Cisco 6500.  Or snmp-server ifindex persist on Cisco routers.  This solution doesn't help servers.