Minor problem: I may have mismanaged renaming a new PV server.
To upgrade an existing PV host to PV 5.2 I had our network folks assign a new physical computer a temporary name in our campus DNS, then ran a clean PVE 5 install on it. I copied over my VM snapshot files, unpacked and ran one, everything seems fine. Next, shut down the old physical machine, and have the network folks assign its name and IP number to my new box. But it wouldn't boot up using the old server's name: I didn't catch at first that the temporary name/IP the new server had pulled with DHCP when I first set it up would get hard-coded into /etc/hosts, /etc/hostname, and /etc/network/interfaces. Corrected those files, now my "new" server looks like the "old" server on the network, NFS mounts to it work, hooray we've succeeded in upgrading.
Almost.... when I run the PVE web console on the new box, I do see the name of my host as one of the nodes. But there is also a node listed with the previous temporary server name. And, the VM I had loaded up when I was testing the server with its temporary name, does not appear on my list of available VMs under the permanent server's node name. (See screenshot. VM is 203. Permanent server name is ee-psi.)
So, my questions: 1) proper way to remove the temporary node and 2) proper way to get my VM on the permanent node.
I can see in /etc/pve/nodes directories named for the temporary server name and the permanent server name. The config file for my VM is in the qemu-server subdirectory of the temporary node. I attempt to simply copy it over to the qemu-server directory of the permanent node. It won't do so.
My Suggested Solutions: I could just reload my VM from the backup snapshots like I did when I first brought it over to the server. I just want to be sure that won't confuse/break something, or that it won't result in using twice as much space because of the "ghost" node. And, I guess I could manually erase the temporary node's directories in /etc/pve/nodes . Again, I don't want to break something. Yeah, I could do another clean PVE install, but I've got things like NFS working and all my snapshots copied (this system is a backup server for another host), would like to avoid that.
To upgrade an existing PV host to PV 5.2 I had our network folks assign a new physical computer a temporary name in our campus DNS, then ran a clean PVE 5 install on it. I copied over my VM snapshot files, unpacked and ran one, everything seems fine. Next, shut down the old physical machine, and have the network folks assign its name and IP number to my new box. But it wouldn't boot up using the old server's name: I didn't catch at first that the temporary name/IP the new server had pulled with DHCP when I first set it up would get hard-coded into /etc/hosts, /etc/hostname, and /etc/network/interfaces. Corrected those files, now my "new" server looks like the "old" server on the network, NFS mounts to it work, hooray we've succeeded in upgrading.
Almost.... when I run the PVE web console on the new box, I do see the name of my host as one of the nodes. But there is also a node listed with the previous temporary server name. And, the VM I had loaded up when I was testing the server with its temporary name, does not appear on my list of available VMs under the permanent server's node name. (See screenshot. VM is 203. Permanent server name is ee-psi.)
So, my questions: 1) proper way to remove the temporary node and 2) proper way to get my VM on the permanent node.
I can see in /etc/pve/nodes directories named for the temporary server name and the permanent server name. The config file for my VM is in the qemu-server subdirectory of the temporary node. I attempt to simply copy it over to the qemu-server directory of the permanent node. It won't do so.
My Suggested Solutions: I could just reload my VM from the backup snapshots like I did when I first brought it over to the server. I just want to be sure that won't confuse/break something, or that it won't result in using twice as much space because of the "ghost" node. And, I guess I could manually erase the temporary node's directories in /etc/pve/nodes . Again, I don't want to break something. Yeah, I could do another clean PVE install, but I've got things like NFS working and all my snapshots copied (this system is a backup server for another host), would like to avoid that.
Last edited:
