A virtual machine on a node in the PVE cluster always shuts down automatically

[runing]

A virtual machine was created on a node in the PVE cluster, but it always shuts down automatically. After deleting the virtual machine, it will still automatically shut down when recreated. The virtual machine deployment method is to use QM create, and the image used is Ubuntu 22.04 LTS. img 。May I ask how to identify the cause of the problem?

Here are the commands to create a virtual machine。
qm create 104 --cores 16 --memory 20480 --net0 virtio,bridge=vmbr0
qm importdisk 104 ubuntu-22.04-LTS.img vm_ssd
qm set 104 --ostype l26 --autostart 1 --ciuser root --cipassword ilove4dbim --scsihw virtio-scsi-pci --scsi0 vm_ssd:vm-104-disk-0 --ide1 vm_ssd:cloudinit --boot c --bootdisk scsi0 --serial0 socket --vga serial0 --ipconfig0 ip=172.30.0.104/24,gw=172.30.0.1 --name ubuntu.3090.pve

 

[Moayad]

Hi,

Did you see anything interesting in the syslog when at the time when the VM shutdown?

 

[gfngfn256]

qm importdisk 104 ubuntu-22.04-LTS.img vm_ssd

Don't see the disk ubuntu-22.04-LTS.img being use-added/attached to the VM afterwards.

As per the docs:

qm disk import <vmid> <source> <storage> [OPTIONS]
Import an external disk image as an unused disk in a VM. The image format has to be supported by qemu-img(1).

 

[runing]

Hi,

Did you see anything interesting in the syslog when at the time when the VM shutdown?

Is it looking at the node's syslog or the virtual machine's syslog? Where can I see this syslog

 

[runing]

Don't see the disk ubuntu-22.04-LTS.img being use-added/attached to the VM afterwards.

As per the docs:

The disk is already in use by the virtual machine and can be seen on the web interface of the PVE cluster

 

[gfngfn256]

The disk is already in use by the virtual machine and can be seen on the web interface of the PVE cluster

Of course it is - but is it attached?

 

[runing]

Of course it is - but is it attached?

This is the hardware configuration of that virtual machine

 

[_gabriel]

too many RAM allocated ?
iirc, PCI passthrough lock RAM allocated.

 

[gfngfn256]

IDK what version of PVE you are using - but under Processor, I see no type.

Try changing Processor Type to host (since it appears you are doing passthrough)

 

[gfngfn256]

IDK what version of PVE you are using - but under Processor, I see no type.

Replying to my own query - I imagine your VM is probably using Default kvm64 type - hence not showing anything.

As my above post try changing Processor Type to host (since it appears you are doing passthrough).

 

[runing]

Replying to my own query - I imagine your VM is probably using Default kvm64 type - hence not showing anything.

As my above post try changing Processor Type to host (since it appears you are doing passthrough).

I'm not sure how to change to Processor Type host. Do you have any relevant reference documents? Thank you

 

[runing]

too many RAM allocated ?
iirc, PCI passthrough lock RAM allocated.

Is it possible that the virtual machine can run normally, but will automatically shut down after running for a period of time, or is it PCI passthrough locking RAM? How to check if PCI passthrough locks RAM? Is there a solution?

 

[gfngfn256]

I'm not sure how to change to Processor Type host. Do you have any relevant reference documents?

Just go to Web GUI, Node, VM, Hardware, Processors, Press Edit & change Type (list through the dropdown). Its really easy.

 

[_gabriel]

How to check if PCI passthrough locks RAM? Is there a solution?

afaik, no
https://forum.proxmox.com/threads/w...-point-driver-locked-usage.139025/post-620796

 

[runing]

Just go to Web GUI, Node, VM, Hardware, Processors, Press Edit & change Type (list through the dropdown). Its really easy.

Is it a CPU? is this one?

 

[gfngfn256]

Is it a CPU? is this one?

Exactly.

 

[runing]

afaik, no
https://forum.proxmox.com/threads/w...-point-driver-locked-usage.139025/post-620796

Thank you very much.

 

[runing]

Exactly.

Changes have been made. Thank you very much