I've noticed an odd issue that seems to crop up from time to time with NPM, and I thought I'd check in with the community and see if anyone else has bumped into this.
All of our NPM/NCM nodes are added manually. We have a mix of ICMP/SNMP nodes and SNMP nodes, the ICMP-only nodes are generally just windows servers, and are there so we can alert if something goes down/up/etc. The SNMP nodes, all Cisco enterprise networking devices with a handful of Meraki's thrown in for good measure, are there so we can poll up/down status as well as all the other good stuff you get out of SNMP (interface stats, note utilization stats, topology, etc).
I have run into problems with the SNMP nodes over and over again. If we swap, let's say a Cisco ISR 4321 with a Meraki MX67, that will generally never be reflected in Solarwinds until I manually run a rediscovery on that node. Is this expected behavior? I would have thought the scheduled rediscovery would do this for us.
I have also run into a recent problem with our Catalyst 2960-X switches. They're all stack capable, but there are a handful that do not show as compatible with the default solarwinds switch stack poller. A rediscovery does not fix this, but if I go into the node and List Resources, make no changes, and submit that...the poller is suddenly compatible with the node and all is well.
So in a nutshell...Should rediscovery handle these things? Does anyone else have to manually rediscover nodes to get their machine type and other info to update? Has anyone else seen a situation where you have to manually List Resources to get something to 'snap' into place?
For the record, we're running 2020.2.6 as of last week.
