NCM Nightly Config Report Errors - Connection Refused

thad   For me the issue was that the device was somehow loosing the NCM profile.  I had to edit the device and change it to the appropriate profile again.

That's strange. I haven't had that issue that I know of... All of the switches I have checked from the list still have their original NCM profiles set. Thanks for the reply

I usually find the following to be true:

  ERROR: Cannot download Startup config : Cannot Log into Device : bad password. 

    Issue: Using telnet when it should be ssh or vice versa

  ERROR: Cannot download Running config : Connection Refused

    Issue: Same as above

  ERROR: Cannot download Running config : Connectivity issues, discarding configuration (or configuration is too short).

    Issue: Using a default template or the Cisco IOS when connecting to a WLAN controller or some other non router/switch device. Try changing the templates to the specific device type

Have you tried to connect to the device from the poller that is monitoring it? There could be an ACL in place that is preventing the connection and that is why you are getting connection refused. If you can ssh to the device from the poller then there is something else going on with the device in NCM. I have had some devices show up as an unknown device type before which caused issues with NCM and I had to remove and re-add them for it to work right again.

jdane07​ I have switched a few switches from Telnet to SSH or vice versa and they still show up on the report the next day. A few that I took note of were not in the report the next day even though I made no changes to them, so that's another strange thing going on.

@Chris T I have changed all of the switch ACLs to include the solar winds IP. However- if I'm on the poller page, nothing happens when I try to SSH from the webpage; & telnet has been hit/miss. Could this be part of the problem, or is just another random issue? I can go to the configs tab of the node and download the config just fine.

It could be another random issue. Have you logged into the physical (or virtual) server that the poller software is on and tested the connection to the device that way? If you can SSH or Telnet from the actual server desktop then the problem is narrowed down to the SolarWinds configuration somewhere.

In addition to the other suggestions given already, I would suggest you enable session tracing to help your troubleshoot the issues.  The trace files can be a little hard to parse initially but basically they are showing you exactly what NCM is doing and how it is interpreting the responses it gets from your devices.  That way you can see in debug level detail exactly why you had issues and try to figure out how you can fix them, if you can't figure it out support is going to ask you to generate traces anyway so it doesn't hurt to have them. One handy trick is to search for the lines with <-- because that shows you where NCM was actually sending data to the device.

See the Success Center article here,  Generate session trace logs

-Marc Netterfield

    Loop1 Systems: SolarWinds Training and Professional Services

•                 LinkedIN: Loop1 Systems
•                 Facebook: Loop1 Systems
•                 Twitter: @Loop1Systems

Thanks for the suggestion, we tested this and we're still able to Telnet/SSH into the specific switch

Hey thanks for the reply, we checked the session trace log and actually found that the config downloader just stopped at 223 nodes left. It didn't show any errors however and just said config download complete. Do you have any idea why it would just stop abruptly?

Can you see if all of the devices that are having errors are on the same poller? If the devices are on the same poller and every devices that is managed in NCM on that poller is having issues it may be an issue with the NCM poller software on that poller. Could be a version miss match between the version on the poller and the version on the Core server, or possibly something else. Not sure what could cause that, just fishing for ideas at this point.