NETGEAR is aware of a growing number of phone and online scams. To learn how to stay safe click here.
Forum Discussion
RBR50 / RBS50 intermittent routing problem
I've had my Orbi setup for about 10 months and, with exception of it being a nightmare to setup (it doesn't play well with Sonos), it's worked fairly reliably. Over the first 8-9 months It needed a r...
mpsamuels wrote:I have DHCP and DNS setup on a server that's connected by ethernet to RBS50A and DHCP setup on the RBR50 that only has a 2 IP address scope with each of those IPs reserved for the two RBS50s. I have the same reservations setup on my server so regardless of which DHCP service responds to the request for an address first, the RBS50s will always have the same IP address and there will never be an IP conflict. I have daisy chaining topology disabled.
This is certainly a creative configuration. Connecting this DHCP/DNS server to a satellite rather than the router is a bit unusual. It would be helpful to understand more about the problem, such as any connection with power outages or device restarts. Can it be replicated consistently?
DHCP requests are broadcast packets, so I would not describe this as "routing traffic that is destined for..." Broadcast packets are supposed to go everywhere.
Netgear's implementation of DHCP does not match my own personal expectations. It appears, for example, that a request for DHCP renewal does not incorporate the MAC reserbation table. So, when a device says, "I want to renew 192.168.1.25, the Orbi does an ARP to verify that no one is also using that IP and says, "sure. go for it." Not bothering to check the reservation table and say, "no bad boy. You are supposed to be on 192.168.1.99" I have no clue what a DHCP server does when it gets a request and has no IP's left in the pool. The Wikipedia article on DHCP does not mention this special case https://en.wikipedia.org/wiki/Dynamic_Host_Configuration_Protocol#Discovery
There's no correlation with power outages and I've no way to replicate the problem at the moment, it seems to happen at random.
It may be a 'creative' solution to attach the server to a satelite rather than the router but given the location of the server and the fixed placement of my WAN modem by my ISP I have no choice. This config has worked almost flawlessly for 9 months and continues to work 90% of the time even now. I see no reason why it shouldn't work either.
I'm aware of how DHCP is meant to work, maybe I should be more specific and clarify that:
1) In the cases a client doesn't get a DHCP address while connected to RBR50, if I give it a static address in the OS and try to ping the IP of either RBS50A, the server or any other device connected to RBS50A I receive no reply. If I ping RBS50B or any WAN IP address I receive responses as expected, hence this is only an issue routing traffic to RBS50A. An initial DHCP request may be a broadcast packet in itself but the complete handshake and allocation of an address does involve the use of routing protocols.
2) Sometimes the problem occures when my client device already has an IP address leased and I experience problems as I have lost connectivity to my DNS server i.e as above while connected to RBR50 and I try to ping the IP of either RBS50A or the server I receive no reply. If I ping RBS50B or any WAN IP address I receive responses as expected
That DHCP implementation does seem a little odd but as I'm not seeing any evidence of IP address conflicts I don't see that as being the cause of the problem.
What happens when RBS50A and RBS50B are switched?
(does the problem go with the satellite?)
Is it always the same device or does this happen randomly to many devices?
I hadn't tried switching the satellites but have just done so for confirmation. However, as I highly suspect it's an issue with the router rather than the satellite, I'm imaginig the problem is likely to return within a day or two.
As for 'is it always the same device' - when it happens, the problem effects every client device on my LAN and it impacts their ability to reach either RBS50A or the devices attached to it. However, it's only really the inability to reach the server providing DHCP and DNS for my LAN (attached to RBS50A) that really causes any grief. Strangely, devices attached to RBS50A can still reach the other devices connected to RBS50A and route outwards to the WAN so these experience no problems whenever devices connected to RBS50B or RBR50 are essentially rendered offline by virtue of not being able to reach their DNS server.