Can not start new created VM

[muekno]

First I am new to Proxmox, but not to Linux or virtualisation, used ESX since more than 20 years.

After having running Proxmox creating vmbr0 I created my first VM using a tutorial I found at LINUX HANDBOOK. It's not a great thing giving a name select the uploaded ISO, I used the Debian bookworm netinstaller I created the server itself, so it should be fine, left the rest by defaults.
Proxmox created the VM without any complains. Now go to _console and tried to start, got an error after seconds, tried some more times, error again,
(connection timed out, start failed QUEMU exited with code 1),

What I am doing wrong
Thanks for help

 

[BobhWasatch]

Maybe post the error message?

 

[muekno]

Maybe post the error message?

I did, perhaps not so clear
"connection timed out" or "start failed QUEMU exited with code 1" depending if I start from console or from the start button above

Regards
Rainer

 

[sb-jw]

Please provide a few details about your setup and the Config qm config VMID for the VM.

 

[muekno]

See Screenshots, may be my bridge config vmbr0 is not ok, but that should not be a matter at that time.

Here is my /etc/network/interfaces :

mk@proxmox:~$ cat /etc/network/interfaces
# network interface settings; autogenerated
# Please do NOT modify this file directly, unless you know what
# you're doing.
#
# If you want to manage parts of the network configuration manually,
# please utilize the 'source' or 'source-directory' directives to do
# so.
# PVE will preserve these directives, but will NOT read its network
# configuration from sourced files, so do not attempt to move any of
# the PVE managed interfaces into external files!

source /etc/network/interfaces.d/*

auto lo
iface lo inet loopback

auto enp2s0f0
iface enp2s0f0 inet static
        address 172.16.1.77/24
        gateway 172.16.1.254
        network 172.16.1.0
        dns-domain muekno.de
        dns-search muekno.de
        dns-search gerdakloos.de
        dns-nameserver mail.gerdakloos.de

iface enp2s0f0 inet6 auto

iface enp2s0f1 inet manual

iface enp3s0f0 inet manual

iface enp3s0f1 inet manual

auto vmbr0
iface vmbr0 inet static
        bridge_ports none
        bridge_stp off
        bridge_fd 0
mk@proxmox:~$

I hope that helps, if you need more, please ask and give a hint where to find

 

[BobhWasatch]

To be clear, when you say "created the VM", did you mean the netinstall was a success and it then failed to reboot, or that the netinstall never was able to start? Reason I ask is because boot order is scsi, then ide. It will only fall back to ide if there is no scsi disk, which in this case there is. So if there is no OS on the scsi disk you need to change to boot order in order to get the installer to run.

 

[sb-jw]

Whether the installation was successful or not or the VM tries to boot from the wrong medium does not actually play a role in the error pattern. Apparently the KVM process itself can't even start, but that has nothing to do with netinstall or the boot medium.

dns-domain muekno.de
dns-search muekno.de
dns-search gerdakloos.de
dns-nameserver mail.gerdakloos.de

I honestly have no idea what exactly you're trying to do with this. I would definitely scrap this and not use it.
In particular, you should specify a DNS server with its IPs, because these are supposed to resolve the name. But if the server first has to resolve the IPs from the DNS server, but doesn't have any IPs from DNS servers so that it can resolve them, how is it supposed to resolve DNS queries?

bridge_ports none

I don't know if QEMU responds to this or not, actually I would say it shouldn't matter. But you can still try removing the network interface and see if it works.

Unfortunately you didn't say anything about your hardware, so I'm guessing that your CPU either doesn't support virtualization or it isn't activated in the BIOS. However, the VM has the “KVM” flag activated, which you should also use.

 

[BobhWasatch]

Whether the installation was successful or not or the VM tries to boot from the wrong medium does not actually play a role in the error pattern. Apparently the KVM process itself can't even start, but that has nothing to do with netinstall or the boot medium.

The install cannot have worked. As you point out, the DNS settings are bogus. Also, the vmbr0 bridge has no connection to anything but is the VM's network device. Didn't notice that before. The proper setup for the bridge is more like this (untested but should be close):

iface enp2s0f0 inet manual

iface enp2s0f1 inet manual

iface enp3s0f0 inet manual

iface enp3s0f1 inet manual

auto vmbr0
iface vmbr0 inet static
        address 172.16.1.77/24
        gateway 172.16.1.254
        dns-domain muekno.de
        dns-search muekno.de
        dns-search gerdakloos.de
        dns-nameserver 172.16.1.254 # or some other IP address of a recursive nameserver
        bridge_ports enp2s0f0
        bridge_stp off
        bridge_fd 0
       
iface vmbr0 inet6 auto

 

[muekno]

To be clear, when you say "created the VM", did you mean the netinstall was a success and it then failed to reboot, or that the netinstall never was able to start? Reason I ask is because boot order is scsi, then ide. It will only fall back to ide if there is no scsi disk, which in this case there is. So if there is no OS on the scsi disk you need to change to boot order in order to get the installer to run.

I mean I have configured the VM, then I try to start it, like I did it on ESX, there i have a path to the ISO Media, confige a disk ram etc. and start it. or just on bare metall I boot from an USB Stick have the bootable ISO on it.

So what I want to do here is the starting the installation.

Whether the installation was successful or not or the VM tries to boot from the wrong medium does not actually play a role in the error pattern. Apparently the KVM process itself can't even start, but that has nothing to do with netinstall or the boot medium.


I honestly have no idea what exactly you're trying to do with this. I would definitely scrap this and not use it.
In particular, you should specify a DNS server with its IPs, because these are supposed to resolve the name. But if the server first has to resolve the IPs from the DNS server, but doesn't have any IPs from DNS servers so that it can resolve them, how is it supposed to resolve DNS queries?


I don't know if QEMU responds to this or not, actually I would say it shouldn't matter. But you can still try removing the network interface and see if it works.

Unfortunately you didn't say anything about your hardware, so I'm guessing that your CPU either doesn't support virtualization or it isn't activated in the BIOS. However, the VM has the “KVM” flag activated, which you should also use.

My Hardware is a HP DL380 G6, single CPU, 146GB HD Raid 1 with spare, 8 GB RAM, should be enough for testing. Does run fine under ESXi 6.5. As I am looking for a VMware alternative I like to test Proxmox.

As I configured my "erste" VM i accepted the defaults, I added just the VMs name and the name of the ISO file. I think Proxmox knows what is does when giving defaults.

 

[BobhWasatch]

Do you have any errors in the syslog (Datacenter > <host> > syslog in the GUI) when you try to start the VM?

ETA: Just saw your other thread about installation problems. Apparently you installed over Debian? Did you follow the guide for doing that? Looks like some steps may have been missed. That would explain why the bridge setup is incorrect. Maybe there's something else?

https://pve.proxmox.com/wiki/Install_Proxmox_VE_on_Debian_12_Bookworm

 

[muekno]

The install cannot have worked. As you point out, the DNS settings are bogus. Also, the vmbr0 bridge has no connection to anything but is the VM's network device. Didn't notice that before. The proper setup for the bridge is more like this (untested but should be close):

iface enp2s0f0 inet manual

iface enp2s0f1 inet manual

iface enp3s0f0 inet manual

iface enp3s0f1 inet manual

auto vmbr0
iface vmbr0 inet static
        address 172.16.1.77/24
        gateway 172.16.1.254
        dns-domain muekno.de
        dns-search muekno.de
        dns-search gerdakloos.de
        dns-nameserver 172.16.1.254 # or some other IP address of a recursive nameserver
        bridge_ports enp2s0f0
        bridge_stp off
        bridge_fd 0
      
iface vmbr0 inet6 auto

Thats correct, just my mistake, I corrected it and DNS resolution works fine now.
But at that time of starting installation of the VM from the ISO file it should not harm.


But I found out that is seams the installation of Proxmox created Certificates with a domain they got from my Fritz Box.
Thats for the Proxmox web GUI itself and for the VNC Console.

I apreciate your help

Regards

Rainer

 

[BobhWasatch]

One other thing...Proxmox default is setting X86-64-v2-AES for CPU type in the VM settings. Apparently some versions of your HP server came with processors that did not support AES instructions. Proxmox default may be a bit optimistic here. You could try setting CPU type to "host" or to "kvm64".

 

[muekno]

Do you have any errors in the syslog (Datacenter > <host> > syslog in the GUI) when you try to start the VM?

ETA: Just saw your other thread about installation problems. Apparently you installed over Debian? Did you follow the guide for doing that? Looks like some steps may have been missed. That would explain why the bridge setup is incorrect. Maybe there's something else?

https://pve.proxmox.com/wiki/Install_Proxmox_VE_on_Debian_12_Bookworm

Yes I followed this URL, I don't think I missed something
Concerning the logs, I just can see the errors in the task on the bottom and if I change to cluster log there are the same

Do you have any errors in the syslog (Datacenter > <host> > syslog in the GUI) when you try to start the VM?

ETA: Just saw your other thread about installation problems. Apparently you installed over Debian? Did you follow the guide for doing that? Looks like some steps may have been missed. That would explain why the bridge setup is incorrect. Maybe there's something else?

https://pve.proxmox.com/wiki/Install_Proxmox_VE_on_Debian_12_Bookworm

I followed this URL and do not think I missed something but there seams somthing wrong with the cerifucates

 

[muekno]

One other thing...Proxmox default is setting X86-64-v2-AES for CPU type in the VM settings. Apparently some versions of your HP server came with processors that did not support AES instructions. Proxmox default may be a bit optimistic here. You could try setting CPU type to "host" or to "kvm64".

That's the trick, set it to host and installation starts.
Thank you very much
Regards
Rainer

 

[sb-jw]

Apparently some versions of your HP server came with processors that did not support AES instructions.

There is no CPU for the system that cannot do AES-NI. The feature is either deactivated or the mainboard revision does not offer it.

As I already mentioned, looking into the BIOS would have been helpful.