Changed Host IP on Network; Can't Get To Gateway Anymore

[wheen]

A few months ago I changed my whole network over from 192.168.1.0/24 to 10.20.30/23. In Proxmox, I changed the hosts IP from the GUI (192.168.1.7 -> 10.20.30.7) and everything worked fine.

Well, today I took the host down (For the first time since changing the network configuration) to add some more RAM to it and when I boot back up it tells me to visit 192.168.1.7:8006 and I can't connect to it.

Being a spaz who sucks at Googling, I couldn't find posts like https://forum.proxmox.com/threads/wrong-welcome-message-after-changing-ip.41290/ immediately, that told me to edit /etc/hosts specifically. Instead I ran a find /etc/ -xdev -type f -print0 | xargs -0 grep -H "192.168.1.7" and manually changed anything I found to 10.20.30.7. This included /etc/hosts, but it also included one or two other locations and I'm not sure if that broke something. Then I rebooted, and this changed the banner at the start, but everything was still broken.

One that I haven't changed yet (it's read only... and chmod as root won't let me change it?) is /etc/pve/corosync.conf which still lists 192.168.1.7

As for the symptoms, I can't connect to anything from within the host. I hooked up a keyboard to it and while logging in, I can't ping the gateway or anything else. checked and double-checked the physical connections and they're all solid.

Any insight into how I could fix this would be much appreciated!

 

[wolfgang]

Hi,

to write at pmxcfs (/etc/pve/*) you need quorum if you lost quorum than you can force write access with

pvecm expected 1

I hooked up a keyboard to it and while logging in, I can't ping the gateway or anything else.

This would indicate you have an network problem.
Check your Network settings in /etc/network/interfaces and your network cables.[/Code]

 

[wheen]

Hi!

Thanks for the response, I'll try the pvecm thing later today. I have a one-node network, so have not really dealt with quorom before. That said, corosync sounds like its cluster-related. Would it still have an effect in a one-node cluster?

All my network cables are fine and double checked, and /etc/network/interfaces seems correct, but I'm not 100% sure what to look for.

I also just remembered something else that may be relevant. At some point, I did try and add another node to the cluster, but it failed and I chose to tackle my problem from another angle. Could something in that process be causing this issue?

 

[wolfgang]

Would it still have an effect in a one-node cluster?

No but you wrote

s /etc/pve/corosync.conf which still lists 192.168.1.7

so I assume you have a cluster because on a one node setup this file do not exist.

Could something in that process be causing this issue?

No but this would explain the corosync file.

Do you have the firewall enabled?
If so disable it and test again.

 

[wheen]

I finally got around to messing with it again.

Firewall is disabled. The physical connection is solid. Here is my /etc/network/interfaces (Copied manually):

auto lo
iface lo inet loopback

iface enp3s0 inet manual

auto vmbr0
iface vmbr0 inet static
    address 10.20.30.7
    netmask 255.255.254.0
    gateway 10.20.30.1
    bridge_ports enp3s0
    bridge_stp off
    bridge_fd 0

Again, I manually edited this. Desired address, netmask, and gateway should all be accurate relative to the rest of the network, but I'm not sure about anything else. Does that look fine?

 

[wolfgang]

Does that look fine?

Yes