And I have links from two ISP (ISP1, ISP2) and Mikrotik router with combo port: SFP and Ethernet ports under single interface “combo1”.
ISP1 cable is plugged to Ethernet port and configured with static IP address (IP → Addresses) on interface “combo1”. DHCP server is supported by ISP1, but static address is required to be used for my config.
ISP2 cable is plugged to SFP port and cannot be configured with static IP address, DHCP client usage is required.
I can configure both static address and DHCP Client for “combo1” interface, so ISP1 and ISP2 will work good when “Compo Mode” (current active combo-port) is switched in “combo1” interface settings. But, when I use ISP1 (configured with static address), enabled DHCP client generates error (IP → DHCP Client → “Status” column). This seems logical: since the IP address is already statically configured for “combo1” interface, DHCP Client cannot do it again. This error occurs over and over again, in a loop, while ISP1 is active in “Combo Mode” (Ethernet port). This seems wrong to me, it shouldn’t be like this, I think.
Is it possible to set up DHCP Client only for one of two compo ports (SFP + ISP2), ot something similar? Or am I worrying for nothing and the client error is not a problem?
One port can have many IP addresses, so the fact that there is already static one doesn’t prevent DHCP client from trying to get another.
Wouldn’t it be better to use two independent ports, one for each ISP? You could use both at the same time, do automatic failover, anything. Using combo port doesn’t seem very flexible.
Yes. But in my case, static IP address for ISP1 and dynamic IP address from ISP1’s DHCP server are equal, this is the cause of the error in the DHCP Client in router.
It is useful where one of ISP’s links has SFP connector (optical fiber). Manually switching the combo port (which ISP should be used now) is more easy, than switching separate ports in “internet bridge”, I think. Combo port is more easy to setup and can do automatic failover without any configs:
combo-mode (auto | copper | sfp; Default: auto)
When auto mode is selected, the port that was first connected will establish the link. In case this link fails, the other port will try to establish a new link.
If both ports are connected at the same time (e.g. after reboot), the priority will be the SFP/SFP+ port.
When sfp mode is selected, the interface will only work through SFP/SFP+ cage.
When copper mode is selected, the interface will only work through RJ45 Ethernet port.
If ISP1 also supports DHCP (together with static IP addresses), then you can simply remove the static IP address assignment.
As to ease of using combo port: as you can see, the ease fades due to lack of configuration flexibility. The automatic failover between the modes seems to rely on loss of link (heartbeat) which doesn’t always happen, specailly so if there’s any kind of CPE involved. Using separate interfaces allows to check L3 (IP) connectivity and failover in case L3 connectivity is lost. Not to mention the rest of goodies already mentikned by @Sob. In that case one doesn’t switch port in “WAN bridge”, one rather manipulates routing tables.
Yes, ISP1 supports DHCP. But for some reasons, IP should be statically configured, I wrote abot this in original post.
In my case, there is enough flexibility (I make a few tests), so combo port usage is looks like a more easy solution. Manual failover configuration for separate ethernet ports and playing with route tables is not so easy, I think.
A physical outage of the local or adjacent interface or the connecting cable (electrical or optical) between them is equally likely at your place as somewhere inside the network of your ISP, but there are simply more places where this can happen in the ISP’s network, plus there are also other possible outages that can happen there. So the overall probability that you will not be able to access the internet via a given ISP is much higher than the probability that the physical connection between your router and the adjacent ISP device breaks.
In another words, using the combo port requires no configuration effort but only covers a small fraction of the outage causes.
I clarify: “ISP outage” in my country currently means:
The provider’s equipment is physically damaged: a mortar mine, an artillery shell, a cruise missile or a kamikaze drone hit the communication line.
The provider’s equipment is de-energized: a mortar mine, an artillery shell, a cruise missile or a kamikaze drone hit the electrical line or power station.
In both cases, the connection disappears completely and for a long time. For days, not hours. So there is no “small fraction of the outage causes”, there is concrete physical damaged or disabled equipment. In this case, combo port and manual switching of combo-ports are the simplest and most relevant topic for me, and complex failover logic is useless there.
