First, a bit of background on my installation and what I was doing when I ran into problems...
I have two machines running Proxmox. They are not clustered. I am running this version of Proxmox:
pve-manager/5.1-41/0b958203 (running kernel: 4.13.13-4-pve)
I have an unexpired community subscription for both machines, though I've been unable to update from the Enterprise repositories for some time. I get this error:
E: The repository 'https://enterprise.proxmox.com/debian/pve stretch Release' does no longer have a Release file.
When I check my subscription status in the Proxmox web GUI, it can successfully connect to https://shop.maurer-it.com to check the license, but I'm told I have an "Invalid Server ID".
These two machines are on the same network. Within each machine, I have a rather elaborate virtual network set up, consisting of a bunch of virtual switches and quite a few VLANs. Some VLANs stay internal to the machine they're on, and others are intended to move traffic between the VMs on these two machines. However, I've been unable to move tagged traffic between machines. I bought a network tap and found that though the traffic is properly tagged if I capture it within the virtual network, it is not tagged when it gets onto the physical wire. As a result, the traffic never makes it to its destination. So, I did some digging on how to configure a physical Linux network interface to support tagged Ethernet frames.
As a result of my research, I ended up installing package "vlan" via apt-get install vlan. I created interface eth0.20, I created a virtual Linux bridge in the Proxmox GUI, I associated that virtual bridge with eth0.20, and I set the appropriate interfaces on my VMs to use the new Linux bridge. Traffic still wasn't moving, and I noticed the virtual Linux bridge was not up, so I rebooted to allow it to come up.
And now on to the problem I'm having...
After rebooting, none of my VMs started automatically, and the web interface was down. I could manually start the VMs that didn't use the new Linux bridge via qm start / pct start, as appropriate. This prompted me to apt-get remove vlan, to remove eth0.20 and the associated Linux bridge from /etc/network/interfaces, and to reboot. I was no better off. I performed an apt-get dist-upgrade and rebooted. Still no better off.
I tried manually restarting pvedaemon and found that it is masked. I then found that the only /usr/bin/pve* file present is /usr/bin/pvecm. All other /usr/bin/pve* files are missing.
So, it seems that some of the core Proxmox files got uninstalled during this process. I tried falling back to the no-subscription repository and doing an apt-get dist-upgrade, but the pve* files did not come back.
Is there any way I can recover this Proxmox installation or at least save the two KVMs I really care about?
I have two machines running Proxmox. They are not clustered. I am running this version of Proxmox:
pve-manager/5.1-41/0b958203 (running kernel: 4.13.13-4-pve)
I have an unexpired community subscription for both machines, though I've been unable to update from the Enterprise repositories for some time. I get this error:
E: The repository 'https://enterprise.proxmox.com/debian/pve stretch Release' does no longer have a Release file.
When I check my subscription status in the Proxmox web GUI, it can successfully connect to https://shop.maurer-it.com to check the license, but I'm told I have an "Invalid Server ID".
These two machines are on the same network. Within each machine, I have a rather elaborate virtual network set up, consisting of a bunch of virtual switches and quite a few VLANs. Some VLANs stay internal to the machine they're on, and others are intended to move traffic between the VMs on these two machines. However, I've been unable to move tagged traffic between machines. I bought a network tap and found that though the traffic is properly tagged if I capture it within the virtual network, it is not tagged when it gets onto the physical wire. As a result, the traffic never makes it to its destination. So, I did some digging on how to configure a physical Linux network interface to support tagged Ethernet frames.
As a result of my research, I ended up installing package "vlan" via apt-get install vlan. I created interface eth0.20, I created a virtual Linux bridge in the Proxmox GUI, I associated that virtual bridge with eth0.20, and I set the appropriate interfaces on my VMs to use the new Linux bridge. Traffic still wasn't moving, and I noticed the virtual Linux bridge was not up, so I rebooted to allow it to come up.
And now on to the problem I'm having...
After rebooting, none of my VMs started automatically, and the web interface was down. I could manually start the VMs that didn't use the new Linux bridge via qm start / pct start, as appropriate. This prompted me to apt-get remove vlan, to remove eth0.20 and the associated Linux bridge from /etc/network/interfaces, and to reboot. I was no better off. I performed an apt-get dist-upgrade and rebooted. Still no better off.
I tried manually restarting pvedaemon and found that it is masked. I then found that the only /usr/bin/pve* file present is /usr/bin/pvecm. All other /usr/bin/pve* files are missing.
So, it seems that some of the core Proxmox files got uninstalled during this process. I tried falling back to the no-subscription repository and doing an apt-get dist-upgrade, but the pve* files did not come back.
Is there any way I can recover this Proxmox installation or at least save the two KVMs I really care about?
