Open kingtrw opened 6 years ago
It seems we encounter "same" problem with Cisco Nexus switches and HPE switchs.
We have a VPC domain on Nexus switches in order to make distributed LACP which seems to cause problem for netdisco discovery.
The networkMap, based on LLDP seems to work Fine with the last Netdisco release but in the port List of the HPE Switchs we can't see any uplink info in the "Connected Node and device" column
Can we provide more information in order to help resolve the problem ?
in recent netdisco releases a better effort is made to manage topologies where either certain parts of the neighbor info differ or where 1 neighbor didn't bother replying to cdp/lldp.
as such my first advice would be to try the most recent netdisco version (don't forget to update snmp::info too).
that being said, even with the latest version we also see this behaviour on links between c9332pq's and fabric interconnects 6248. the 6248 side show correct neighbor info, the 9332pq side just shows possible uplink.
our 6248 devices are also showing up under the admin menu -> undiscovered neighbors for some reason. do you also see this behaviour?
My "main" edge network is Extreme XOS 22.2.1.5 devices. In some cases, I have an HPE 5.20 Release 1519P03 device hanging off, to provide more edge ports on the default VLAN.
If I look at the 'ports' information for an Extreme device (EXTREMESW), it is listed correctly -- but if I look at the same information for the connected HPE device, it incorrectly displays a neighbour inaccessible icon, then the MAC address & hostname of a server switch (several hops away), followed by the inter-switch communication (required for MLAG) IP address of the OTHER server switch.
View of EXTREMESWITCH:
View of HPESWITCH:
I have checked that LLDP is enabled on HPESWITCH and it is, and it can see EXTREMESWITCH listed as a neighbour in the GUI.
This behaviour is consistent on all 9 of my HPE switches.
I am using NetDisco 2.39.30.