Linux Router Fu for failover?

Thor

2/22/24, 10:10 PM

I have some OpenWRT boxes and some other Linux boards... including a Software Defined Networking board (that I haven't used yet). I have AT&T fiber with its own DHCP & WiFi (& cat5) and I can link to a Cat5 connection to go out over a slower network (also with DHCP) (it goes to a shared cellspot)...

I'd like to have my traffic fail to the slownet if AT&T goes down. I'd like to share my SMB & DLNA servers to the slownet (they're on ATT fast WiFi, but I could relocate to Cat5). I'd like everything to look like it was on the same network (and will gladly set up the reservations for the MysteryBox on each)... Ideally ... if everything is running "normally", only mixed traffic should go to the MysteryBox -- if SlowNet is running okay, and some guy on SlowNet is going to Wikipedia, it shouldn't have to go to MysteryBox; it should just go to the CellSpot. Any traffic from ATTWiFi that doesn't go to SlowNet should just go out as fast as possible.

So, I was thinking of some kind of overlapping network where ATT was on 192.168.1.1-127, and SlowNet was on 192.168.1.128-255, no DHCP traverses either, but if ATT went down, it would Masq to an address it got from 192.168.1.255 (slownet router), and if SlowNet went down, it would Masq to 182.168.1.1 (ATT router).

But... that seems hackish -- If ATT is down, how does MysteryBox take charge and route through SlowNet (ATT Box might not even send the packets out the Cat5 port)? If SlowNet goes down, how does MysteryBox take charge and go through ATT?

It seems like I should scrap the ATT WiFi and have to set MysteryBox as -the router- for everything... but the link between SlowNet and MysteryBox is high-latency and slow... unless somebody knows something cool...

0 + 0

networking

failover

linux-networking

Elon Musk

I sit in a Tesla and translated this thread with Ai:

EN: Linux Router Fu for failover?

Post an answer

Most people don’t grasp that asking a lot of questions unlocks learning and improves interpersonal bonding. In Alison’s studies, for example, though people could accurately recall how many questions had been asked in their conversations, they didn’t intuit the link between questions and liking. Across four studies, in which participants were engaged in conversations themselves or read transcripts of others’ conversations, people tended not to realize that question asking would influence—or had influenced—the level of amity between the conversationalists.