Network Discovery not allow me to add discovered nodes

There's a lot of potential things that I could see stopping this.

Do you get any messages when you try to add them via discovery or do they just not get added to the discovery for import?

Do you have multiple APEs or just the one server?

Is there anything in the logs (C:\ProgramData\SolarWinds\Discovery Engine) that indicate any errors?

I'm assuming you've added the SNMP credentials to your discovery, do you have it set to prefer SNMP and add the nodes via ICMP if it fails?

Thank you for the quick response!

  1.  Do you get any messages when you try to add them via discovery or do they just not get added to the discovery for import?

•   They just don’t show up. Typically I only see 1 or 2 but no interfaces

  2.  Do you have multiple APEs or just the one server?

•   Single server that is separate from the network they are sitting on. Everything is routed to the proper network though.

  3.  Is there anything in the logs (C:\ProgramData\SolarWinds\Discovery Engine) that indicate any errors?

•   Only this for almost every IP in my list to add.

•   2019-12-03 11:09:43,717  DEBUG PollerBase - Poller 'N.Details.SNMP.F5' is not applicable for this node. Exception: SolarWinds.Orion.Pollers.Framework.SNMP.SnmpResponseException: Error 31002 - No Such Name

  4.  I'm assuming you've added the SNMP credentials to your discovery, do you have it set to prefer SNMP and add the nodes via ICMP if it fails?

•   SNMP Creds are correct. I do have it setup as you described.

OK, so to clarify when you run the discovery for the nodes, only a few show up as discovered? If you try increasing the SNMP timeout/retries or removing the SNMP credentials do you get different results?

One of the things that I'm testing/fishing for is to see if they might somehow be caught up in Discovery's deduplication, or if it's a general SNMP failure for some reason.

Success Center

If the MAC, SysName or DNS entries match in some way it may just be dropping them from the discovery. Theoretically, removing the SNMP credentials would see them all added to help confirm this (though it's not the end result it confirms) and then you can follow the above link to disable the deduplication. Otherwise it may be an SNMP thing, but that's an alternate line of troubleshooting.

That is correct and I have tired maxing out the Timeout fields. I have not tried removing the SNMP Creds though. I’ll give that a shot…

I have found the duplication article that talked about that and turned it off in the DB. It had no effect.

Tried it with NO snmp credentials and it returned them all!

Not sure what I would check next, but at least if they're added it should be simple enough to edit them all together and update the polling method and credentials from there.

I also just found out with this vendor (Eltek), they have problems with SolarWinds NPM with detecting their devices via SNMP V2c. I has to change all the devices to V1 to in order for discovery to work. ¯\_(ツ)_/¯

I have not had discoveries fail the discover part unless there was a problem with the credentials, firewalls/acls, or poller issues.  I have had occasional problems with imports, where the discover worked fine, selected what the import including interfaces, and the import claimed to be successful but the nodes were not there after.  We still are not sure what is causing that intermittently, but rebooting the polling engine has fixed it even though no polling engine errors were reported.