Problem trying to recover VM

[erneiball17]

Hi guyus!

After a dist-upgrade the GUI won't work anymore. The synthoms are very weird cause if you try to check the VM's and the CT's there are no existing anymore, if you try to install proxmox-ve you receive the message:

The following packages have unmet dependencies: librados2-perl : Depends: perlapi-5.20.0 but it is not installable lxc-pve : Depends: libgnutls-deb0-28 (>= 3.3.0) but it is not installable pve-cluster : Depends: librrd4 (>= 1.4~rc2) but it is not installable Depends: perlapi-5.20.2 but it is not installable Depends: corosync-pve (>= 2.3.4-1) but it is not going to be installed pve-qemu-kvm : Depends: libgnutls-deb0-28 (>= 3.3.0) but it is not installable Depends: libnettle4 (>= 2.6) but it is not installable Depends: libpng12-0 (>= 1.2.13-4) but it is not installable

And there's no way to fix those packages. But the weird thing is when you enter systemctl status 'pve*' this is the output:

● pve-firewall.service - LSB: Proxmox VE firewall Loaded: loaded (/etc/init.d/pve-firewall; generated) Active: active (exited) since Wed 2021-11-24 15:38:53 -03; 6s ago Docs: man:systemd-sysv-generator(8) Process: 2139 ExecStart=/etc/init.d/pve-firewall start (code=exited, status=0/SUCCESS) nov 24 15:38:53 proxmox1 systemd[1]: Starting LSB: Proxmox VE firewall... nov 24 15:38:53 proxmox1 systemd[1]: Started LSB: Proxmox VE firewall. ● pve-cluster.service - LSB: Starts the pve cluster filesystem Loaded: loaded (/etc/init.d/pve-cluster; generated) Active: active (exited) since Wed 2021-11-24 15:38:53 -03; 6s ago Docs: man:systemd-sysv-generator(8) Process: 2126 ExecStart=/etc/init.d/pve-cluster start (code=exited, status=0/SUCCESS) nov 24 15:38:53 proxmox1 systemd[1]: Starting LSB: Starts the pve cluster filesystem... nov 24 15:38:53 proxmox1 systemd[1]: Started LSB: Starts the pve cluster filesystem. ● pvedaemon.service - LSB: PVE Daemon Loaded: loaded (/etc/init.d/pvedaemon; generated) Active: active (exited) since Wed 2021-11-24 15:38:53 -03; 6s ago Docs: man:systemd-sysv-generator(8) Process: 2127 ExecStart=/etc/init.d/pvedaemon start (code=exited, status=0/SUCCESS) nov 24 15:38:53 proxmox1 systemd[1]: Starting LSB: PVE Daemon... nov 24 15:38:53 proxmox1 systemd[1]: Started LSB: PVE Daemon. ● pvestatd.service - LSB: PVE Status Daemon Loaded: loaded (/etc/init.d/pvestatd; generated) Active: active (exited) since Wed 2021-11-24 15:38:53 -03; 6s ago Docs: man:systemd-sysv-generator(8) Process: 2134 ExecStart=/etc/init.d/pvestatd start (code=exited, status=0/SUCCESS) nov 24 15:38:53 proxmox1 systemd[1]: Starting LSB: PVE Status Daemon... nov 24 15:38:53 proxmox1 systemd[1]: Started LSB: PVE Status Daemon. ● pvefw-logger.service - LSB: PVE firewall logger Loaded: loaded (/etc/init.d/pvefw-logger; generated) Active: active (exited) since Wed 2021-11-24 15:38:53 -03; 6s ago Docs: man:systemd-sysv-generator(8) Process: 2128 ExecStart=/etc/init.d/pvefw-logger start (code=exited, status=0/SUCCESS) nov 24 15:38:53 proxmox1 systemd[1]: Starting LSB: PVE firewall logger... nov 24 15:38:53 proxmox1 systemd[1]: Started LSB: PVE firewall logger. ● pve-manager.service - LSB: PVE VM Manager Loaded: loaded (/etc/init.d/pve-manager; generated) Active: active (exited) since Wed 2021-11-24 15:38:53 -03; 6s ago Docs: man:systemd-sysv-generator(8) Process: 2142 ExecStart=/etc/init.d/pve-manager start (code=exited, status=0/SUCCESS) nov 24 15:38:53 proxmox1 systemd[1]: Starting LSB: PVE VM Manager... nov 24 15:38:53 proxmox1 systemd[1]: Started LSB: PVE VM Manager.

Well after a few attemps I'm not able to recover the GUI but in /dev/pve/ you can see that the vm-ID-disk-1 are still in the server. My question is: It is possible to take those images and migrate themselves to other proxmox and create a new VM and replace the disk? If try to do it I receive the message: mv: error writing '/dev/pve/vm-100-disk-0': No space left on device Also I tryed qemu-img convert vm-100-disk-0 -O qcow2 /dev/pve/vm-100-disk-0 or qemu-img convert vm-100-disk-0 -O qcow2 /dev/pve/vm-100-disk-0.qcow2 and qemu-img convert vm-100-disk-0 -O raw /dev/pve/vm-100-disk-0.raw or qemu-img convert vm-100-disk-0 -O raw /dev/pve/vm-100-disk-0 and replacing them in /dev/pve/qemu-server/100.conf but the error is always the same:

kvm: -drive file=/dev/pve/vm-161-disk-2.qcow2,if=none,id=drive-scsi0,format=raw,cache=none,aio=io_uring,detect-zeroes=on: Could not open '/dev/pve/vm-161-disk-2.qcow2': filesystem does not support O_DIRECT TASK ERROR: start failed: QEMU exited with code 1

I hope you could help me guys because this vm is very important. Thanks to all

 

[Moayad]

Hello,

Please post the output of the following between [CODE][/CODE]

apt update && apt dist-upgrade
grep '' /etc/apt/sources.list
grep '' /etc/apt/sources.list.d/*

Did you remove the proxmox-ve? what says your node if you did the pveversion command?

 

[erneiball17]

Hello Moayad,

Thank you so much for answer. Let me let you know that the incident is fixed. We had to set the pve-no subs repo, validate the keys and start the aptitude install proxmox-ve and it packages. After that we did vgcfgrestore pve --forceand the GUI was able to access again. Finally we have created the vms and replace the disks for the olders in etc/pve/qemu-server/VMID.conf

By the moment the VM is running normally and I did a backup if the problem persists in other time.

Thank you again, have a nive day!