routers running v6.10
One dhcp-client and the other router that is dhcp-server hands out an fixed IP address to the client. Existing situation for weeks, months or even year. Both routers have seen several updates of software or occurred power cycles before. Server-client situations have been normally working for long time without single problem…
But; In the last half year (so also on previous versions of OS) I have had a couple of times where certain clients that ran with a certain handed IP address for long time without any issues, suddenly couldn’t have traffic any more. This happened just like that. No interference for my part nor the client. The only thing that might have happened was a wireless disconnect due power failure or just power down on behalf of client.
And than the problem is there:
No more internet for the client!
In investigating I found that; IP was requested by client, offered by server and accepted by client but still no IP traffic possible. Client’s IP shows in neighbor list of adjacent routers but can’t be pinged. Nor can I ping it over the wireless link from within the registration table.
Other instance was that client that had certain IP leased for months, if not years, suddenly wasn’t handed the same IP any longer…
Any other IP made no problem, but that specific one assigned to client not! Client still requested, server still ‘offered’ but IP was never set anymore. No matter what I tried… IP’s just one step higher or lower made no problem…
(But since all my clients are limited by simple queues based upon their IP address, like my whole admin is based upon, I don’t like IP’s being changed too much..)
In both instances, disabling the original fixed IP in the server for that client made the client immediately request a new, server immediately offered new IP (sometimes the same!) and client accepted! If it was not the same, I could erase the old lease rule and set the new one again with the new IP and it worked!
So I got the original situaion back! IP is the same, mac address is the same, client is not even touched so all the same… WHAT HAPPENED? Why does it work now?
Has anybody seen similar events?
I really don’t understand why this happens. It happens very rare, and random. I can’t repeat the problem but have it now 3 or 4 times with different clients in a 500+ clients network.
It happened with a client with simple routed network. It happened with client where the dhcp-client was actually on a bridge, and now it happened with a client that has to reach server through a vlan tunnel.
But other clients, side by side in the same physical and IP network are not having such problem.
It is like sometime ROS just doesn’t like a certain IP-mac combination in a certain memory allocation.
I can remember that the first time the issue was actually solved in a reboot of my main gateway that also performs the dhcp-server function. But I don’t want to do that often, it disrupts all traffic to all clients and if something goes wrong I’m in big problems…
There must be some explanation… 