I'm a long-time casual linux user, but very new to proxmox and playing with networks. I think this issue relates to Proxmox, but I'm happy to be pointed somewhere else if someone recognizes this problem as relating to some other part of my setup. I'm also happy to share config files or terminal output, but at this point I'm not clear what I should share.
A few months ago I stood up my first Proxmox system, and have slowly been familiarizing myself creating VMs and containers etc. Initially I had no trouble with getting full internet on VMs, however, a month or two ago that changed. The host system and lxc containers DO still have internet with no trouble, however VMs can only connect using IP addresses. Since then I've been poking at things trying to troubleshoot, and I think i have tracked the issue down specifically to DNS for the VMs. Here's the list of my setup and observations from my amateur troubleshooting:
The closest scenario I've found in my searching is this thread: https://forum.proxmox.com/threads/cant-access-internet-from-vms-in-bridge-network.162106/ but My understanding is that I should be able to connect VMs through the linux bridge, and no solution for this was actually presented in that post. Based on all of the above my assumptions/guesses are:
Since the VM can ping with IP but not name, I assume this is something to do with my DNS configuration.
Since other devices on my network are unaffected, I assume the problem is with the configuration of the networking in Proxmox, not with DNS configuration at my router.
Since this affects VM's immediately, before I change any settings, I assume the issue is somewhere in my host configuration, not in the VM's.
Since the host and containers are unaffected but VM's are, I am guessing this is something to do with the bridge interface.
I'm hoping that some fine person(s) might be able to help me troubleshoot.
Thanks
A few months ago I stood up my first Proxmox system, and have slowly been familiarizing myself creating VMs and containers etc. Initially I had no trouble with getting full internet on VMs, however, a month or two ago that changed. The host system and lxc containers DO still have internet with no trouble, however VMs can only connect using IP addresses. Since then I've been poking at things trying to troubleshoot, and I think i have tracked the issue down specifically to DNS for the VMs. Here's the list of my setup and observations from my amateur troubleshooting:
- VMs used to function normally, but are now affected. I have a TrueNAS Scale VM that used to work fine, but now cannot update because apt fails to resolve the update servers
- Any new VM I create is affected by the same issue whether that is one of the standard VMs from the Proxmox interface, or one created manually from an ISO
- The issue affects both server & GUI VMs, I have primarily been using a full GUI Debian 13 VM to troubleshoot.
- The Proxmox host and lxc containers are unaffected
- Other computers on my network are unaffected.
- Turning off the Proxmox firewall does not resolve the issue.
- Setting the firewall "input policy" for the affected VM to 'Accept' does not resolve the issue
- Affected VMs can use IP addresses ping the host (192.168.1.30), the gateway router (192.168.1.1), and the internet (8.8.8.8)
- Affected VMs can NOT use domain names like google.com to ping
- Apps like apt from the command line, or firefox in the GUI return connection errors
- Firefox on a Debian VM is able to open the web management page of o a TrueNas VM using the local IP in the search bar
- Using Debian's advanced network configuration in a VM, I tried explicitly entering DNS IPs for My gateway, unbound running on my pi-hole, and google's DNS, but this does not resolve the issue.
- VMs are all configured to use 'Linux Bridge' network interface with vmbr0
- My router is currently set up to just inherit the DNS from my ISP
- I only have a single network segment. (Gateway is at 192.168.1.1, affected VMs at 192.168.1.xxx)
- I use my router to reserve IP's for all devices VMs containers and the host, so the VMs are usng DHCP to obtain an IP, but the router always supplies the same IP to each VM.
- possibly unrelated: Around the time this occurred I had set up a pi-hole lxc using community scripts, and installed unbound. At the time, I left unbound not configured and did not point any of my systems to use the pi-hole as I was mostly playing with how creation of VMs an lxcs work. I didn't think about this when I noticed these network issues, and now I can't recall iexactly when the issue appeared relative to creating that unbound instance. I have since finished configuring unbound and tried pointing an affected VM at the unbound DNS using Debian's advanced network configuration GUI, but got no luck
The closest scenario I've found in my searching is this thread: https://forum.proxmox.com/threads/cant-access-internet-from-vms-in-bridge-network.162106/ but My understanding is that I should be able to connect VMs through the linux bridge, and no solution for this was actually presented in that post. Based on all of the above my assumptions/guesses are:
Since the VM can ping with IP but not name, I assume this is something to do with my DNS configuration.
Since other devices on my network are unaffected, I assume the problem is with the configuration of the networking in Proxmox, not with DNS configuration at my router.
Since this affects VM's immediately, before I change any settings, I assume the issue is somewhere in my host configuration, not in the VM's.
Since the host and containers are unaffected but VM's are, I am guessing this is something to do with the bridge interface.
I'm hoping that some fine person(s) might be able to help me troubleshoot.
Thanks