DHCP Client lease expired

RouterOS General
1

Good afternoon
I have two providers on ports 1 and 2. I receive an address from both via dhcp. On the second provider I constantly have an error in the logs:

dhcp-client on ether2-beeline lost IP address - lease expired

The address rent is valid only for 10 minutes. Every 10 minutes I lose connection for 1 second because I receive the address at the very last second. Release and Renew buttons don’t have any effect, there is no address update until it expires.
The other day one provider crashed and I’ve noticed that the address began to update immediately. And even the Release and Renew buttons started working.
As soon as the second provider restored the line, everything repeats again.
Writing the address manually is not an option. In this case Internet stops working completely.
Ros version 7.11.2. I think there is a bug in Mikrotik. But maybe I’m wrong and there is some solution?

2

Just guessing but this could be a case where the second provider sees no traffic, hence releases the address early on his end.

What if you open a terminal and ping the gateway on the provider end each 30-60 seconds using that interface. Not your own ip. The gateway on the other end.
To keep the connection alive so to speak.
If that works, you can make it a netwatch function so it gets done automatically.

3

No, I use this channel all the time. Several home clients and one virtual server work through it. So there is constant traffic exchange there.

4

Then maybe you should talk to that provider.

5

As I said at the beginning, if I disable the first provider, the problem disappears.

6

Maybe you have put those ports in a bridge? make sure each of them is not member of any bridge.

7

No, both ports are not included in the bridge. They are only part of the WAN group.

Conducted an experiment. I made the default route for the second provider. The problem is gone.

I didn’t plan to use the second provider in this way, but if it doesn’t work any other way, I’ll live with it.

8

When you have more than one DHCP client of course you should set the default route distance differently in each of them, and probably setup two route tables and a script to insert a default route in the second table, and routing rules or mangling.
I would not think it would cause the issue that you see now when you do not do that, but it could be that your ISP does some monitoring of the connection (e.g. ping) and of course that only works with the properly setup policy routing.