I've recently purchased a Vigor 2865. I connected an old Netgear router to it via ethernet to act as an additional WiFi AP at the other end of the house.

I'm experiencing a strange issue. If I connect a device to the Netgear WiFi I can ping and access any device on the 2865 LAN. If I then connect to the 2865 WiFi the device becomes unable to ping local addresses (apart from the 2865) for about 5 minutes. WAN access is fine. Switching back to the Netgear ping returns. If the device is connected to the 2865 via ethernet within that 5 minute period it works.

All devices and the 2865 WiFi are on LAN1/VLAN0. The Netgear has a static address in the same subnet as LAN1 and DHCP is disabled. I've also tested with a Plusnet Hub One.

I've had the same setup using several different routers over the years without issue. Any ideas why this is causing a problem with the 2865? Thanks!

Is this an ARP cache issue?

As in the 2865 has recorded the Netgear's MAC address for the IP address of the device in question, rather than that of the device itself?

(You would think, though, that if the 2865 thinks it's dealing with two different MAC addresses, there would be two different IP addresses involved).

It might give you a pointer, if nothing else :wink:

Thanks, I think it is an ARP issue, but I don't really know much about ARP. As far as I can see the entries in the ARP cache table show the correct IP and MAC combinations. However I've just noticed if I set my iPad or phone to use a randomised MAC for the 2865 SSID they stop being affected by the issue.

I guess the question is why is the 2865 seemingly unable to cope with what I'd have thought would be a fairly common setup, when cheapo ISP routers can? Is there something "enterprisey" about ARP handling in the 2865 or is it a bug? I do have the ax variant which has only just been released.

Tomek wrote:
If I then connect to the 2865 WiFi the device becomes unable to ping local addresses (apart from the 2865) for about 5 minutes.

I think I intially read this the other way round...

Are you able to see the ARP & Routing tables on the device in question?

Perhaps it is trying to route via the Netgear, rather than ARP for a MAC address and talk directly. I can't think of a mechanism for this happening though...at least not without it being specifically configured somewhere

Are you able to compare a 'traceroute' from the device (when it's working!) to another LAN device (using both Netgear & 2865).

In both cases, of course, the 'pings' should go directly from device-to-device, with no mention of intermediate nodes.