F5F Stay Refreshed Power Users Networks Options for network redundancy Available configurations to ensure reliable connectivity.

Options for network redundancy Available configurations to ensure reliable connectivity.

Options for network redundancy Available configurations to ensure reliable connectivity.

L
LoreleiBella
Junior Member
Find
46
06-27-2016, 01:12 PM
#1
Hello everyone, I recently faced a network disruption due to a failed port on my primary switch. It was a Zyxel multi-gigabit device at the heart of my setup, with 10G ports linking my main server (NAS/PLEX/SQL/Blue Iris/MQTT) and my office desktop. The switch connected to my router and each of my other switches via 1G or 10/100Mbps (the latter for my home automation gear). I've attached a simplified network diagram.

During my absence last weekend, the port on my router fell silent, leaving me reliant only on my wireless connection. To avoid similar issues, I'm exploring redundancy strategies. My main server featured an integrated 2.5G port, so connecting it to a 1G switch seems feasible. My encode server uses a 1G port connected through a PCIe card, which I can link to a 1G switch. The Virtualization Server has four 1G ports and aims to connect two to each 10G and 1G switch. I'm also considering routing all switches back to the router's built-in 4-port 1G port.

I understand the router remains a single point of failure, but this should improve reliability. I plan to test these configurations once my replacement arrives. In the meantime, here are my questions: Both my main server and encode server run Windows 10 Pro. Since I don’t manage switches, link aggregation isn’t an option. Is there a way for Windows to automatically switch to the slower connection only when the primary fails? I don’t want it to attempt transferring a large file over the slower link. Can the system detect a failure if the switch connection is intact but the WAN link is down? My Virtualization server uses Proxmox and runs on VMbr0. Should I add all physical ports to my vmbr0 bridge? Would that cause issues if the four ports connect to two different switches? Am I at risk of switch loops? This is more about best practices. Currently, I access everything via IP addresses, but with multiple IPs expected, how should I reach them? Should I use local DNS or another method?

I’m confident this is supported by pfSense. My router is an Asus RT-AC68U, which I’d like to keep for its built-in DDNS and OpenVPN capabilities. I attempted setting up my own OVPN server but couldn’t get it working. Would a managed switch be preferable here? What budget-friendly or home-friendly 10G managed switches are available—preferably with at least two 10G ports and a couple of 2.5G ports?

My only managed switch experience comes from Rockwell Stratix models, which have a steep learning curve. Thanks!
Reply
Reply
L
LoreleiBella
06-27-2016, 01:12 PM #1

Hello everyone, I recently faced a network disruption due to a failed port on my primary switch. It was a Zyxel multi-gigabit device at the heart of my setup, with 10G ports linking my main server (NAS/PLEX/SQL/Blue Iris/MQTT) and my office desktop. The switch connected to my router and each of my other switches via 1G or 10/100Mbps (the latter for my home automation gear). I've attached a simplified network diagram.

During my absence last weekend, the port on my router fell silent, leaving me reliant only on my wireless connection. To avoid similar issues, I'm exploring redundancy strategies. My main server featured an integrated 2.5G port, so connecting it to a 1G switch seems feasible. My encode server uses a 1G port connected through a PCIe card, which I can link to a 1G switch. The Virtualization Server has four 1G ports and aims to connect two to each 10G and 1G switch. I'm also considering routing all switches back to the router's built-in 4-port 1G port.

I understand the router remains a single point of failure, but this should improve reliability. I plan to test these configurations once my replacement arrives. In the meantime, here are my questions: Both my main server and encode server run Windows 10 Pro. Since I don’t manage switches, link aggregation isn’t an option. Is there a way for Windows to automatically switch to the slower connection only when the primary fails? I don’t want it to attempt transferring a large file over the slower link. Can the system detect a failure if the switch connection is intact but the WAN link is down? My Virtualization server uses Proxmox and runs on VMbr0. Should I add all physical ports to my vmbr0 bridge? Would that cause issues if the four ports connect to two different switches? Am I at risk of switch loops? This is more about best practices. Currently, I access everything via IP addresses, but with multiple IPs expected, how should I reach them? Should I use local DNS or another method?

I’m confident this is supported by pfSense. My router is an Asus RT-AC68U, which I’d like to keep for its built-in DDNS and OpenVPN capabilities. I attempted setting up my own OVPN server but couldn’t get it working. Would a managed switch be preferable here? What budget-friendly or home-friendly 10G managed switches are available—preferably with at least two 10G ports and a couple of 2.5G ports?

My only managed switch experience comes from Rockwell Stratix models, which have a steep learning curve. Thanks!

Reply
Reply