VM showing a disk from other VM?

[mcfly9]

I ran into a very weird issue on Proxmox 8.1.3.

Have two Windows servers, running fine for a few years now. I wanted to add a new disk into each. I added the disk using the UI to each of the VM's onto the SCSI single controller with disk number 2 (scsi2). To my surprise, in both of the VM's, two new drives showed up. I have added a serial to the disks to be able to distinguish between the disks, and again, to my surprise, it looks like both disks are attached to both VM's.
If I shut down one of the VM's, the "foreign" disk disappears from the running VM.

What have I messed up? Time for starting fresh and doing a proxmox reinstall?



VM303:

agent: 1
bios: ovmf
boot: order=virtio0;ide2
cores: 4
cpu: host
efidisk0: VMDisk-WD:vm-303-disk-0-efi,efitype=4m,pre-enrolled-keys=1,size=4M
ide2: none,media=cdrom
machine: pc-q35-8.1
memory: 32768
meta: creation-qemu=7.1.0,ctime=1674926458
name: exch-g-03
net0: virtio=AA:8B:6F:1A:0F:38,bridge=vmbr0
numa: 0
onboot: 1
ostype: win11
scsi2: VMDisk-WD:vm-303-disk-2,discard=on,iothread=1,size=255G,ssd=1,serial=1234
scsihw: virtio-scsi-single
smbios1: uuid=ec5f528e-be6d-404f-bd7f-c5c4beb18a74
sockets: 2
startup: order=3,up=30
virtio0: VMDisk-WD:vm-303-disk-1-os,discard=on,iothread=1,size=127G
virtio1: VMDisk-WD:vm-303-disk-3-mbxold,discard=on,iothread=1,size=256G
vmgenid: f3be573e-802c-4608-9e8d-c2f6f9ad79bf

VM304:

agent: 1
bios: ovmf
boot: order=virtio0;ide2;net0
cores: 4
cpu: host
efidisk0: VMDisk-Samsung:vm-304-disk-0-efi,efitype=4m,pre-enrolled-keys=1,size=4M
ide2: none,media=cdrom
machine: pc-q35-8.1
memory: 32768
meta: creation-qemu=7.1.0,ctime=1674926458
name: exch-g-04
net0: virtio=42:85:54:5D:2A:4D,bridge=vmbr0
numa: 0
onboot: 1
ostype: win11
scsi2: VMDisk-Samsung:vm-304-disk-2,discard=on,iothread=1,size=255G,ssd=1,serial=2345
scsihw: virtio-scsi-single
smbios1: uuid=9f327560-19a6-4291-9b87-5afa2d54591a
sockets: 2
startup: order=3,up=30
virtio0: VMDisk-Samsung:vm-304-disk-1-os,discard=on,iothread=1,size=127G
virtio1: VMDisk-Samsung:vm-304-disk-3-mbxold,discard=on,iothread=1,size=256G
vmgenid: caa156bc-4f4b-4e12-bce4-4508bc3b8b96

Note: I have previously renamed the other disks using lvrename while the VM's were powered off and modified the VM conf with the right suffix.

 

[mcfly9]

If I only boot one of the VM's, it correctly shows up:



As soon as I boot up the second VM, the second disk shows up too, albeit without a disk number:

 

[mcfly9]

If I add a new disk using virtio, only the correct disks show up. So the issue seems to be limited to SCSI single.

 

[bbgeek17]

What is the output of "ls -al /etc/pve" and "ls -al /etc/pve/qemu-server/" ?

Its unclear what exactly you renamed and whether there were any land mines left behind. I'd first restart the services :
"systemctl try-reload-or-restart pvedaemon pveproxy pvestatd pvescheduler" and/or reboot the host for good measure.

When you have an issue, capture the output of "ps -efwww|grep [vmid]" - does the wrong disk appear there?
Or is it possible that the VM you are logging into is not the one you think it should be?

In summary, what you are describing should not be possible under normal system operation. Also review the log portion of when the disk is being created, ie after adding the disk "journalctl -n 100"

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[bbgeek17]

This should make the output of "ps" a bit more readable:

ps -efww|grep "id 3002"|sed 's/ -/\n -/g'

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[mcfly9]

What is the output of "ls -al /etc/pve" and "ls -al /etc/pve/qemu-server/" ?

root@proxmox:~# ls -al /etc/pve
total 18
drwxr-xr-x   2 root www-data     0 Jan  1  1970 .
drwxr-xr-x 108 root root     12288 Jan 10 14:28 ..
-rw-r-----   1 root www-data   451 Jan 17 18:22 authkey.pub
-rw-r-----   1 root www-data   451 Jan 17 18:22 authkey.pub.old
-r--r-----   1 root www-data  9778 Jan  1  1970 .clusterlog
-rw-r-----   1 root www-data    36 Jan  6 14:03 datacenter.cfg
-rw-r-----   1 root www-data     2 Jan  1  1970 .debug
-rw-r-----   1 root www-data   495 Jan 10 15:47 domains.cfg
drwxr-xr-x   2 root www-data     0 Nov 25  2022 firewall
drwxr-xr-x   2 root www-data     0 Oct  8  2098 ha
-rw-r-----   1 root www-data   144 Jan 10 15:34 jobs.cfg
lrwxr-xr-x   1 root www-data     0 Jan  1  1970 local -> nodes/proxmox
lrwxr-xr-x   1 root www-data     0 Jan  1  1970 lxc -> nodes/proxmox/lxc
drwxr-xr-x   2 root www-data     0 Jul 17  2023 mapping
-r--r-----   1 root www-data    40 Jan  1  1970 .members
drwxr-xr-x   2 root www-data     0 Oct  8  2098 nodes
lrwxr-xr-x   1 root www-data     0 Jan  1  1970 openvz -> nodes/proxmox/openvz
drwx------   2 root www-data     0 Oct  8  2098 priv
-rw-r-----   1 root www-data  2078 Oct  8  2098 pve-root-ca.pem
-rw-r-----   1 root www-data  1675 Oct  8  2098 pve-www.key
lrwxr-xr-x   1 root www-data     0 Jan  1  1970 qemu-server -> nodes/proxmox/qemu-server
-rw-r-----   1 root www-data     0 May 12  2023 replication.cfg
-r--r-----   1 root www-data  2828 Jan  1  1970 .rrd
drwxr-xr-x   2 root www-data     0 Oct  8  2098 sdn
-rw-r-----   1 root www-data   112 Jan 29  2023 status.cfg
-rw-r-----   1 root www-data   463 Jul  5  2023 storage.cfg
-rw-r-----   1 root www-data   391 Jan 18 16:00 user.cfg
-r--r-----   1 root www-data   996 Jan  1  1970 .version
drwxr-xr-x   2 root www-data     0 Oct  8  2098 virtual-guest
-r--r-----   1 root www-data   998 Jan  1  1970 .vmlist
-rw-r-----   1 root www-data   120 May 12  2023 vzdump.cron
root@proxmox:~# ls -al /etc/pve/qemu-server
lrwxr-xr-x 1 root www-data 0 Jan  1  1970 /etc/pve/qemu-server -> nodes/proxmox/qemu-server
root@proxmox:~#

Its unclear what exactly you renamed and whether there were any land mines left behind.

I have used lvrename to append a suffix to the disk names and after that updated the VM conf. No renames that would wire the disks across VM's.

vm-304-disk-1 -> vm-304-disk-1-os

Will do a clean boot now and try to repro the issue.

 

[bbgeek17]

"ls -al /etc/pve/qemu-server/"

the command requested had a trailing dash that you omitted.

---

Blockbridge : Ultra low latency all-NVME shared storage for Proxmox - https://www.blockbridge.com/proxmox

 

[mcfly9]

Before I waste too much of your time, let me check one thing first and report back.

This might be Windows, in combination with the Exchange server on it thinking that this is some sort of a network attached storage and showing the disk from the other computer (DAG cluster) and potentially nothing to do with Proxmox.