This looks similar to https://forum.proxmox.com/threads/turned-off-vm-still-respond-to-ping-and-ssh.25396/ but I am running "pve-manager/5.4-5/c6fdb264 (running kernel: 4.15.18-12-pve)" .
Short description: a ping to the machine is still working even after shutting it down, but "ssh" is not working.
Context:
There was a CT that still had Wheezy on it and as it lacked monitoring, the sql server that has stopped was not restarted.
I upgraded to jessie and then stretch. During the process I had diffuculties after a "poweroff" or "shutdown" launched from within the CT. Performing 'pve shutdown' and 'pve stop' did not help either - I ended up killing the related lxc process(es).
After that I could still not start the CT on the physical server, so I migrated it to the backup server where it was able to run.
The CT's have an IP on the VPN visible amongst the physical machines.
Now, after having run properly for some time (a few days), the web server inside the CT is not responding to the proxy running on the main machine. It turned out that the web server was reachable on the physical machine it was running on, but not from the main machine. On both physical machines the ping worked, and after stopping the server, the ping still worked.
So I guess that something is still responding to the IP address on the main server, but no other physical machine as the said IP.
The resolve the issue in the short term, I have assigned another IP to the CT and reconfigured the proxy, with the original IP still being pingable.
Maybe somebody has a suggestion what I can do to identify the reason for that, and avoid this in the future.
Short description: a ping to the machine is still working even after shutting it down, but "ssh" is not working.
Context:
There was a CT that still had Wheezy on it and as it lacked monitoring, the sql server that has stopped was not restarted.
I upgraded to jessie and then stretch. During the process I had diffuculties after a "poweroff" or "shutdown" launched from within the CT. Performing 'pve shutdown' and 'pve stop' did not help either - I ended up killing the related lxc process(es).
After that I could still not start the CT on the physical server, so I migrated it to the backup server where it was able to run.
The CT's have an IP on the VPN visible amongst the physical machines.
Now, after having run properly for some time (a few days), the web server inside the CT is not responding to the proxy running on the main machine. It turned out that the web server was reachable on the physical machine it was running on, but not from the main machine. On both physical machines the ping worked, and after stopping the server, the ping still worked.
So I guess that something is still responding to the IP address on the main server, but no other physical machine as the said IP.
The resolve the issue in the short term, I have assigned another IP to the CT and reconfigured the proxy, with the original IP still being pingable.
Maybe somebody has a suggestion what I can do to identify the reason for that, and avoid this in the future.