Virtual Knot Not Running

[Thomas_H]

Hi proxmox'rs,

My Name is Thomas and I am really new to proxmox and virtuality. I have have knowledge in linux and I can say I am not an expert on linux, but I know how to handle the terminal and how to be lazy. Now I have situation I've never had:

I rebooted the hardwareserver, without shutting down the one and only virtual machine, named mailcleaner.
After the reboot, the proxmox-knot was avail, but mailcleaner was still grey, the configuration was wrong and I always got a "timed out", when I tried to start the container.

While googling I found the 'pve'-command and tried to start my virtual machine with the number 100 with a

pve start 100

I got an error like this:

Configuration file 'nodes/proxmox/qemu-server/100.conf' does not exist

okay, no problem. Then do a symbolic link from nodes/mailcleaner/qemu-server/100.conf to nodes/proxmox/qemu-server - but it failed!

root@proxmox:/etc/pve/nodes/proxmox/qemu-server# ln -s /etc/pve/nodes/mailcleaner/qemu-server/100.conf /etc/pve/nodes/proxmox/qemu-server/
ln: failed to create symbolic link ‘/etc/pve/nodes/proxmox/qemu-server/100.conf’: Function not implemented

okay... then copy it.

root@proxmox:/etc/pve/nodes/proxmox/qemu-server# cp /etc/pve/nodes/mailcleaner/qemu-server/100.conf /etc/pve/nodes/proxmox/qemu-server/
cp: cannot create regular file ‘/etc/pve/nodes/proxmox/qemu-server/100.conf’: File exists

huh? O.O qm says, file does NOT exist, bash says it exists? Okay, an ls will show it (hopefully):

root@proxmox:/etc/pve/nodes/proxmox/qemu-server# ls -a
.  ..  101.conf

Okay, seems like file-system is corrupted, but it's all clean...

Any ideas, what happened? Or can tell me anybody, where I was silly?

At least the machine runs again, but on 101.conf The 100 is still there, but grey.

Thanx a lot.

Thomas_H

Can anyone explain this?

 

[walterced]

Search image VM, create now and copy it

 

[dietmar]

huh? O.O qm says, file does NOT exist, bash says it exists? Okay, an ls will show it (hopefully):

Files in /etc/pve/ are stored inside a distributed, cluster wide database. This is not a POSIX file system
(no support to create symbolic links), and has some protection to avoid stupid things. For example
you cannot use the same VMID on another host (that is why your cp failed).

see https://pve.proxmox.com/wiki/Proxmox_Cluster_File_System_(pmxcfs)

 

[dietmar]

If you want to steal a VM from another node, simply use 'mv' instead of 'cp'. For example:

# mv /etc/pve/nodes/mailcleaner/qemu-server/100.conf /etc/pve/nodes/proxmox/qemu-server/

 

[Thomas_H]

Hi Dietmar and walterced,

thanx for your replies. The problem is not any more, that I can't run the machine (I am lucky), because I "stealed" the 100.conf from the mailcleaner-machine and cp'd it to the proxmox-machine. Mailcleaner is running again, but I want to fix that error, which is still there. And it's important for me to have a clean-running system.

How can I do that? Is there a doku or some words I can search for?

Thanx,

Thomas_H

 

[dietmar]

Sorry, I am a bit confused. Seems 'mailcleaner' is a node name (or a virtual machine? Or a container?). How many nodes do
have in the cluster?

 

[Thomas_H]

Hi dietmar,

for sure, I use the wrong terms, so I've added an image of my actual proxmox.

As I have understood proxmox, I have a container, named mailcleaner, in this container should run a VM named 100. By copying the 100.conf to the proxmox-container, I also have a VM named "101", which is running well, but due to hand-copied conf-file to "101" it uses the same ressources like "100".

I guess, what I can do now is the following:

1) shutdown the VM named 101
2) clone the VM named 101
3) delete the container "mailcleaner" and the VM inserted named "100"
4) delete the VM 101
5) start the cloned VM

Is that right?

Thanx for your help...

Thomas_H

 

[Thomas_H]

Okay... I was really breave today and started "to clean" my proxmox. I shutted down the virtual machine, and tried to delete the virtual machine named "100" with the web-frontend. It failed. Then I connected via ssh with the server and moved the /etc/pve/nodes/mailcleaner to /tmp, just for testing. Then I restarted the server and reconnected to the web-frontend. The container "mailcleaner" was away and the web-frontend seems to be tidy and clean. The virtual machine "102" is still running fine. I hope I have done all what have to be done. I still have the directory in "/tmp", so I am able to have a look into the files, if I have to remove a lil bit more than only the container-dir.

And... seems like I have produced the whole error for myself. I've never worked with LVM-volumes, and I mounted a /dev/sdb1 to the proxmox-machine and used it for backups. When updateting and restarting the hardwareserver so the mount of /dev/sdb1 was not there...

If you have any tips for me, what I also should do to have a clean system, let me know.

Thanks in advance...

Thomas_H