Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!


SoYouStart & Failover IPs Proxmox VE 6
New on LowEndTalk? Please Register and read our Community Rules.

All new Registrations are manually reviewed and approved, so a short delay after registration may occur before your account becomes active.

SoYouStart & Failover IPs Proxmox VE 6

Hi all,

Looking for some guidance if possible, I've done various searches on google/ovh docs but still hitting a bit of a brick wall.

where I cam so far:

using Proxmox VE 6.1-5
I have 4 Failover IPS
I've created a MAC address for one of the address's I'm wishing to use and have added this mac address to the network adapter of the VM in proxmox.
I've then logged into the VM and checked my network devices. ip link show - i have lo and ens18
I've then logged into the VM (Debian 10) and edited the /etc/network/interfaces
I've used the following config (recomended in: )

auto lo eth0
iface lo inet loopback
iface eth0 inet static
address FAILOVER_IP
netmask 255.255.255.255
broadcast FAILOVER_IP
post-up route add GATEWAY_IP dev eth0
post-up route add default gw GATEWAY_IP
pre-down route del GATEWAY_IP dev eth0
pre-down route del default gw GATEWAY_IP

eth0 = I've replaced with ens18
FAILOVER_IP = the failover ip associated to the mac address i created
GATEWAY_IP = the main ip of the server replacing the end with .254

at the moment I just get failed to start Raise network interfaces and can't bring up ens18 :(

Thank you in advance for any help you can provide.

Comments

  • Any error messages? Anything in your logs?

    Tried this? https://pve.proxmox.com/wiki/OVH

  • @PainlessHosting said:
    Any error messages? Anything in your logs?

    Tried this? https://pve.proxmox.com/wiki/OVH

    In the VM I get the failed error to bring up ens18, that looks similar to the Ovh guide I used but some of config is worded differently.

    I will try and report back.

  • how does your bridge config on the node look like?

  • I'm assuming you restarted the network service or rebooted before attempting to bring the interface up.

    The link is up/connected in Proxmox, right? Doesn't have "disconnected" checked in the VM options? Have you tried leaving it as eth0? (shouldn't matter if ens18 is in your device list, but worth a shot).

    Also, as as Falzo said, share your node network config and maybe a screenshot of the options page in the PVE GUI for the VM in question.

  • dfroedfroe Member, Host Rep

    The net-tools package which includes the old ifconfig and route commands has been deprecated since Debian 9. You should no longer use them as they are probably not installed by default anymore.

    Most likely bringing up the ethernet device will fail because your post-up commands are failing.

    I am wondering why so complicated to manually add the default route via a post-up route command. I know it was requiered in earlier times to handle a default gateway outside your own subnet (note the /32 netmask) but this isn't the case anymore.

    At least with Debian 10 something much easier like this should work:

    auto ens18
    iface ens18 inet static
    address FAILOVER_IP
    netmask 255.255.255.255
    gateway GATEWAY_IP
    
  • dfroe said: At least with Debian 10 something much easier like this should work:

    auto ens18
    iface ens18 inet static
    address FAILOVER_IP
    netmask 255.255.255.255
    gateway GATEWAY_IP
    

    probably want to add

    pointopoint GATEWAY_IP

    to that ;-)

    btw: does Buster use ifupdown oob at all or is it already switched to netplan? can't remember to have it installed from scratch, so not sure what is used...

    Thanked by 2dfroe citrix
  • dfroedfroe Member, Host Rep
    edited December 2019

    @Falzo said: probably want to add pointopoint GATEWAY_IP to that ;-)

    Indeed, that's a good "point". ;)

    btw: does Buster use ifupdown oob at all or is it already switched to netplan?

    AFAIK it is still ifupdown. And I hope they will keep it as it usually just works. I don't have the need for yet another super fancy Python-based magic hosted on a .io domain. :D
    I love my Debian because it is some kind of old school. The other folks are welcome to switch to Ubuntu. :D

    Thanked by 1Falzo
  • hi all,

    @Falzo / @dfroe / @PainlessHosting

    Here are the details below.

    on the node (main box) the network/interface file contains the following.

    this is the network parts on the gui...

    Many Thanks

    BTW i will try some of the suggestions above :) my example config was taken from the recomendation from OVH knowledge base here: https://docs.ovh.com/gb/en/dedicated/network-bridging/

    Thanked by 1Falzo
  • Marcel40625Marcel40625 Member
    edited December 2019

    Dont forget about

    "post-up echo 1 > /proc/sys/net/ipv4/ip_forward"

    in the Host Network config!

  • @Falzo said:

    dfroe said: At least with Debian 10 something much easier like this should work:

    auto ens18
    > iface ens18 inet static
    > address FAILOVER_IP
    > netmask 255.255.255.255
    > gateway GATEWAY_IP
    > 

    probably want to add

    pointopoint GATEWAY_IP

    to that ;-)

    btw: does Buster use ifupdown oob at all or is it already switched to netplan? can't remember to have it installed from scratch, so not sure what is used...

    i have applied this :) now i can ping the vm and connect to it via ssh using the failover ip, but i cant seem to get internet access when on the box, cant ping / wget / apt-get etc.

  • edited December 2019

    Can you ping a public IP (1.1.1.1)? If so, sounds like a DNS issue, edit /etc/network/interfaces with...

    dns-nameservers 1.1.1.1 8.8.8.8

    Or whatever DNS you prefer.

    (Edited: resolve.conf is a dynamic file managed by the resolveconf service, disregard my previous suggestion. I'm out of practice with Debian)

    Thanked by 1citrix
  • citrixcitrix Member
    edited December 2019

    @PainlessHosting said:
    Can you ping a public IP (1.1.1.1)? If so, sounds like a DNS issue, edit /etc/network/interfaces with...

    dns-nameservers 1.1.1.1 8.8.8.8

    Or whatever DNS you prefer.

    (Edited: resolve.conf is a dynamic file managed by the resolveconf service, disregard my previous suggestion. I'm out of practice with Debian)

    ok it looks like i can ping ip address's i've pinged some other dedicated boxes i have and 1.1.1.1 and i get responses.

    hmmmmm, so possible something to do with DNS?

    i did add dns-nameservers 1.1.1.1 1.0.0.1 to the interface and rebooted but no change.

    EDIT: think i have worked it out resolvconf was not installed and no resolv.conf

Sign In or Register to comment.