I am following the current documentation to setup an hyperconverged network with proxmox. I was previously using a mesh setup, but since I wanted a full redundant network I figured the setup described on this wkik (translated in english) would do the trick:
https://translate.google.com/transl...xmox:hyperconverged-proxmox-2-liens&sandbox=1
But how I could expose the access to the VMS to the public? Any ida whaat need to be configured on the router? Any hint is welcome
This is an abstract of the link above:
The solution is based on BGP, routing in peer to peer mode, VxLAN, CEPH.
It should be understood that we are going to realize that it is not native to Proxmox and will ask to bypass certain interface constraints by configuring them in CLI.
The objective will be to create this infrastructure:
with the following specificities:
https://translate.google.com/transl...xmox:hyperconverged-proxmox-2-liens&sandbox=1
But how I could expose the access to the VMS to the public? Any ida whaat need to be configured on the router? Any hint is welcome
This is an abstract of the link above:
The solution is based on BGP, routing in peer to peer mode, VxLAN, CEPH.
It should be understood that we are going to realize that it is not native to Proxmox and will ask to bypass certain interface constraints by configuring them in CLI.
The objective will be to create this infrastructure:
with the following specificities:
- The IPs in the 10.50.0.0 and 10.43.0.0 network are routed in eBGP.
- The IPs of the 10.50.0.0 network are used for CEPH links, and are routed in priority on the physical DATA network
- The IPs of the 10.43.0.0 network are used for links between VMs, and are routed in priority on the physical VM network. The IP in this network should be used as “router-id”.
- If the DATA network is unreachable, the node will go through the VM network, and vice versa.
- The IPs for the corosync rings are routed statically.
