10 Replies Latest reply on Nov 13, 2012 11:05 PM by byrona

    Problem adding a node

    byrona

      I have the beta appliance up and running and I have added the agent to a Windows node on the same network; however, that node is not appearing in my "Manage Nodes" screen in the LEM interface.

        • Re: Problem adding a node
          byrona

          I have tried uninstalling, rebooting and reinstalling the agent and it still isn't showing up in the manage nodes screen on the LEM appliance.  Is there a newer version of the agent that I should be using?

            • Re: Problem adding a node
              nicole pauls

              Hey Byron,

               

              5.3.1 is the latest agent (there's a hotfix after that, but it's mostly performance related). We only release agent upgrades if necessary and try to keep them to a minimum so people aren't constantly pushing updates.

               

              If it's not appearing in Manage>Nodes, chances are good it can't connect to the appliance. There's a log on the agent in the install directory (which is likely C:\windows\system32 (or syswow64)\ContegoSPOP\) that should give you more info. The other possibility is you're out of licenses, but if you just installed, that shouldn't be the case.

               

              Let me know if you spot anything in the logs.

                • Re: Problem adding a node
                  byrona

                  The install log indicates Successes = 124, Warnings = 0, NonFatalErrors = 0, Fatal Errors = 0, Action Notes = 0.

                   

                  The license on the appliance indicates that it expires on Dec 8th 2012.

                    • Re: Problem adding a node
                      nicole pauls

                      One step further - C:\Windows\SysWOW64\ContegoSPOP\spoplog.txt (on the agent, in case that wasn't implied)

                        • Re: Problem adding a node
                          byrona

                          Ah, in there it clearly shows that it can't connect to the manager though I have no idea why it can't.  The manager is on the same LAN with no firewall in the mix.

                            • Re: Problem adding a node
                              nicole pauls

                              You could try this:

                              Stop the Agent service

                              Remove the c:\windows\syswow64\contegoSPOP\spop directory

                              Start the Agent service

                               

                              This will clear any cached certificates and force it to try again.

                               

                              If it still can't connect, the initial connect (cert exchange) is over 37890, then the follow-up connect is 37892. The agent uses a fixed pool of ports as well, 37893-37896 by default (so, 37893-6 on agent end connects to 37890-2 on appliance end).

                               

                              Windows firewall?

                                • Re: Problem adding a node
                                  byrona

                                  The windows firewall is completely turned off as that was my first suspect.

                                   

                                  I tried the procedure that you suggested and it still seems like it is having a problem based on the following log (IP has been removed to protect the guilty)...

                                   

                                  (Mon Nov 12 10:53:44 PST 2012) WW:WARNING [NioComNetworkParent v24745] {ComModuleSpop:20} Failed to bind socket to local port in range: 37893 - 37896 or to connect to manager at *.*.*.*/37890 for message 0;

                                    • Re: Problem adding a node
                                      nicole pauls

                                      Sometimes AV blocks the connect attempts, too. Anything else already bound to 37893-37896? (seems pretty unlikely, but worth a shot) You could try a telnet from the agent system to the manager IP on 37890 and see if it times out, picks it up, or "other"

                                       

                                      You might also check the SolarWinds Alerts (or Overview > LEM Internal Events) filter to see if the manager side is noting anything about that agent - if it blocks the attempt for some reason, it'll be noted with an event. That should be pretty unlikely since you cleared up the certs and we allow duplicate agents with the same name to connect, but worth checking.

                                       

                                      I checked for any KBs, but it looks like I've exhausted the same steps here (e.g. SolarWinds Knowledge Base :: Troubleshooting LEM Agent Connections, 64 bit).