[SOLVED] How to delete a VM or Container that has no storage and no or missing storage pool.

[francoisd]

Hi,

I have an old VM that was existing on an old pool that I stopped and deleted in between.
The VM is still listed in the web GUI, but I can't delete it.
From the GUI, I have the error :
TASK ERROR: storage 'pool1' does not exist
(which is true). And from the command line:

root@hystou2:~# grep 118 /etc/pve/.vmlist
"118": { "node": "hystou3", "type": "lxc", "version": 76 },
root@hystou2:~# qm destroy 118
Configuration file 'nodes/hystou2/qemu-server/118.conf' does not exist

 

[francoisd]

Just realized, since it's an LXC, qm might not be the best way ;-)

So :

root@hystou2:~# lxc-destroy --name=118
root@hystou2:~# lxc-destroy --name=118
lxc-destroy: 118: tools/lxc_destroy.c: main: 242 Container is not defined
root@hystou2:~# grep 118 /etc/pve/.vmlist
"118": { "node": "hystou3", "type": "lxc", "version": 76 },

As you can see, it seems to delete the container on the first execution since there's no error, and a "Container is not defined" during the second execution. But it remains in the GUI and in the /etc/pve/.vmlist

So, how should I definitely delete traces of this container ?

 

[ph0x]

Is there a 108.conf in /etc/pve/qemu-server? If there is no other VM 118 you probably should delete that file.

 

[francoisd]

You mean 118.conf, but no, there's none in qemu-server. However, thanks for this double check, since I found the definition of the container in :
/etc/pve/lxc/118.conf

So, the lxc-destroy did not remove it.

After digging a bit more in the docs, the Proxmox way to manage lxc is with the pct command:

root@hystou3:~# pct destroy 118
storage 'pool1' does not exist

At least, the error message is consistent with the one from the Web GUI.
The --force gives the same answer:

root@hystou3:~# pct destroy 118 --force
storage 'pool1' does not exist

 

[ph0x]

First thing I would do is setup a bogus pool named pool1 and try again.
The second thing is deleting 118.conf

 

[francoisd]

Hi,

root@hystou3:~# pvesm add dir pool1 --path /mnt/pool1
root@hystou3:~# pvesm list pool1
Volid Format  Type      Size VMID
root@hystou3:~# pct destroy 118
unable to parse directory volume name 'vm-118-disk-0'
root@hystou3:~# pct destroy 118 --force 1 --purge 1
unable to parse directory volume name 'vm-118-disk-0'

Well, this did not really help. So, I finally root@hystou3:~# mv /etc/pve/lxc/118.conf /root/Archive/ which almost instantly removed the LXC 118 from the web interface.

Was not really the way I was expecting, but it seems fine.

Thanks ph0x,

 

[mikeinnyc]

Try this next time:
lxc-ls --fancy
lxc-destroy {-n name} [-f] [-s]

root@NODE1:~# lxc-ls --fancy
NAME STATE   AUTOSTART GROUPS IPV4        IPV6 UNPRIVILEGED
101  RUNNING 0         -      10.10.1.101 -    true         
103  STOPPED 0         -      -           -    true         
104  STOPPED 0         -      -           -    true         
201  STOPPED 0         -      -           -    true         
302  STOPPED 0         -      -           -    true         
303  STOPPED 0         -      -           -    true         
304  STOPPED 0         -      -           -    true         
305  STOPPED 0         -      -           -    true         
306  STOPPED 0         -      -           -    true         
root@NODE1:~# lxc-destroy -n 103 [-f] [-s]
root@NODE1:~# lxc-destroy -n 104 [-f] [-s]
root@NODE1:~# lxc-destroy -n 105 [-f] [-s]
lxc-destroy: 105: ../src/lxc/tools/lxc_destroy.c: main: 240 Container is not defined
root@NODE1:~# lxc-destroy -n 201 [-f] [-s]
root@NODE1:~# lxc-destroy -n 202 [-f] [-s]
lxc-destroy: 202: ../src/lxc/tools/lxc_destroy.c: main: 240 Container is not defined
root@NODE1:~# lxc-destroy -n 302 [-f] [-s]
root@NODE1:~# lxc-destroy -n 304 [-f] [-s]
root@NODE1:~# lxc-destroy -n 305 [-f] [-s]
root@NODE1:~# lxc-destroy -n 306 [-f] [-s]
root@NODE1:~# lxc-ls --fancy
NAME STATE   AUTOSTART GROUPS IPV4        IPV6 UNPRIVILEGED
101  RUNNING 0         -      10.10.1.101 -    true         
303  STOPPED 0         -      -           -    true         
root@NODE1:~# lxc-destroy -n 303 [-f] [-s]
root@NODE1:~# lxc-ls --fancy
NAME STATE   AUTOSTART GROUPS IPV4        IPV6 UNPRIVILEGED
101  RUNNING 0         -      10.10.1.101 -    true         
root@NODE1:~#

 

[mikeinnyc]

root@NODE1:/etc/pve/local# cd qemu-server

root@NODE1:/etc/pve/local/qemu-server# ls

100.conf  400.conf  901.conf  903.conf  905.conf  907.conf

200.conf  900.conf  902.conf  904.conf  906.conf

root@NODE1:/etc/pve/local/qemu-server# rm 100.conf

root@NODE1:/etc/pve/local/qemu-server# rm 200.conf

root@NODE1:/etc/pve/local/qemu-server# rm 400.conf

root@NODE1:/etc/pve/local/qemu-server# rm 900.conf

root@NODE1:/etc/pve/local/qemu-server# rm 901.conf

root@NODE1:/etc/pve/local/qemu-server# rm 902.conf

root@NODE1:/etc/pve/local/qemu-server# rm 903.conf

root@NODE1:/etc/pve/local/qemu-server# rm 904.conf

root@NODE1:/etc/pve/local/qemu-server# rm 905.conf

root@NODE1:/etc/pve/local/qemu-server# rm 906.conf

root@NODE1:/etc/pve/local/qemu-server# rm 907.conf

root@NODE1:/etc/pve/local/qemu-server#

 

[phocks]

I solved this by:

1. edit /etc/pve/lxc/100.conf or whatever number your container is.
2. find the line rootfs: usb_thin:vm-111-disk-0,size=16G.
3. edit usb_thin or whatever the dead drive was called to a storage pool that exists.
4. then pct destroy 100 works or in the Proxmox GUI destroy works too.

I guess just check that the new storage pool doesn't have a disk with the same name, otherwise it may delete that disk.

 

[deckardpj]

I commented out the rootfs: with a # then it was poosible to delete the container from the gui