Proxmox VE GUI Inaccessible

[sch1vv]

Hello. Gone down quite the rabbit hole of troubleshooting a GUI not presenting and am unsure how to proceed.

The setup: proxmox VE running on an AMD ryzen 1600x processor from a few years ago. VM running on top of that hosting a few basic services, set to auto boot. Accessing the GUI previously worked over https://192.168.1.100:8006, but does not connect anymore.

I was able to operate as normal until a few hours ago when the GUI randomly cut out. I have attempted to troubleshoot the GUI not presenting and have deduced that the host device is unable to even be pinged by another computer. Somewhat surprisingly, the VM can still be pinged and even ssh'd into, but the host has almost no connectivity.

I isolated the proxmox server and a laptop on a router with firewall effectively completely disabled. The proxmox server can ping my laptop, but my laptop cannot ping the server. Logs in /var/log/pveproxy end a few hours before the issue began, even over reboots of the system (the last line is "GET / HTTP/1.1" 301- . I have followed this to disable iptables for the time-being, but it changes nothing. I have restarted both pveproxy and pvedaemon services multiple times and both seem to be operating normally. ip a shows the ip that corresponds to the data in /etc/hosts.

Where can I go from here? I am relatively new to linux and am stuck on trouble-shooting at this point. Thanks.

 

[bbgeek17]

The first thing to check is the output of "ip a".
Is CARRIER link detected? Is the NIC in UP state? Is the IP what you expected?
If physical layer is ok, then, I would check for duplicate IPs. Compare the ARP output of the PVE host to what other hosts see.

If you cant ping between IPs, then application layer does not matter (ie service restarts). Until you can ping to your GW and other hosts on the same LAN, you wont get anywhere.

Go back and try to recall what you have changed. Did you modify DHCP ranges? Modified your switch/router?

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[sch1vv]

Thanks for the quick response. Output of "ip a" shows the loopback and the physical layer are up. A quick check with "$ cat /sys/class/net/enp0s25/carrier" when connected to the isolated router also confirms this. As mentioned, I can ping from the proxmox machine to a laptop, but not the other way around. arp output on both the laptop and the proxmox machine show the same ip addresses.

Here is the output on my laptop:

arp -a

Interface: 192.168.1.157 --- 0x13
  Internet Address      Physical Address      Type
  192.168.1.1           7c-8b-ca-4f-e7-58     dynamic
  192.168.1.100         e0-d5-5e-26-c4-d0     dynamic
  192.168.1.255         ff-ff-ff-ff-ff-ff     static
  224.0.0.22            01-00-5e-00-00-16     static
  224.0.0.251           01-00-5e-00-00-fb     static
  224.0.0.252           01-00-5e-00-00-fc     static
  239.255.255.250       01-00-5e-7f-ff-fa     static
  255.255.255.255       ff-ff-ff-ff-ff-ff     static

Interface: 169.254.83.107 --- 0x26
  Internet Address      Physical Address      Type
  169.254.255.255                             static
  224.0.0.22                                  static
  224.0.0.251                                 static
  224.0.0.252                                 static
  239.255.255.250                             static

ping 192.168.1.100

Pinging 192.168.1.100 with 32 bytes of data:
Request timed out.

I believe this began when adjusting parameters for the vm. I stopped the vm and made adjustments to the cpu allocation, and the GUI stopped responding. Inspecting /etc/pve/qemu-server/100.conf shows that the changes I made to the cpu allocation did not save and are as they were before alterations were made. I am unsure why this would result in the GUI not being accessible. No other changes were made to my home router, but to confirm I set everything up on a separate router and network for testing. The laptop and proxmox machine are the only devices connected (no WAN connection for the router).

 

[bbgeek17]

Is there a duplicate IP? Can you ping/ssh from VM to PVE and/or laptop?
Can you shutdown the VM and change ICMP outcome?
Connect the laptop directly into PVE, does that change the ICMP outcome?

You can run below command to confirm that PVE services up, your issue is at lower level:
curl -sk https://localhost:8006
curl -sk https://192.168.1.100:8006

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[sch1vv]

Running the curl commands echos html code, which I assume indicates that the services are up. There are no duplicate ips on the network. I shut down the vm and ensured that no other vms were running, this did not change the ability to ping the pve. Connecting directly to the pve also does not change the ability to ping. Obviously it is a little challenging to paste the output from the pve directly here, but I can provide images:

 

[bbgeek17]

there is no need for html screenshot, your statement that you see it is sufficient to establish that PVE services are ok.
The output of "ip a" would be more interesting. I'd look at tcpdump/wireshark.
Perhaps you have one of those Realtek cards that have been having issues with latest Kernel:
https://forum.proxmox.com/threads/p...t-connection-not-the-typical-solutions.87606/

Its interesting that you seem to have one way communication, perhaps points to some sort of misconfigured firewall on your windows client.

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[sch1vv]

Despite changing nothing, I am now able to connect to the GUI. Attached is the output of ip a and a check for the realtek card. Is there some recommended driver update that needs to be installed? Apologies for sounding naive here, but this has happened a few times with this setup (the cycle of unable to access -> change nothing -> able to access -> change nothing -> unable, etc) so I would love to take any preventative steps if possible.

 

[bbgeek17]

The screenshot clearly shows a NO-CARRIER/DOWN on the interface. Thats a physical layer and/or chipset problem. May be your cable is frail and works one way only.

I have not been keeping tabs on Realtek drama, you will have to search/google on your own.

Good luck

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[sch1vv]

Following up here: I believe the device showing no carrier is the wifi, not the ethernet link that everything is physically connected by. I also no longer see the GUI (as mentioned, been in this loop before lol) and am getting a "connection reset" warning from the browser. Still confused as to where to go :/

 

[Kungolf]

It seems that the wifi interface is "wlp8s0". But IMHO you should instead try to bring up the interfaces "enp5s0" (probably physical) and after that the bridge "vmbr0" using command:

# ip link set dev <interface> up

please note that the IP mask for vmbr0 is apparently 0 (zero), which probably is not what you want.
I hope it helps a bit...

 

[bbgeek17]

llowing up here: I believe the device showing no carrier is the wifi,

You are mistaken. The default PVE bridge is "vmbr0" - it shows "NO/CARRIER".
The interface with index2 - enp5s0 also shows "No/CARRIER". Additionally it shows its master is "vmbr0". The naming of this interface is usually reserved for non-wifi NICs.
The wlp8s0 is commonly wifi interface, it also shows "DOWN".

Essentially no ports in the config are up. And given that NIC is Realtek, I'd search the forums for more information.

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox