[SOLVED] LVM Volume of VM are missig after VM is stopped on V5.4.3

[feldi]

After Upgrading Proxmox to Version 5.4.3

The Virtual LVM Disk (/dev/vg01/vm-xxxx_disk-1) are only visible if the virtual maschine xxxx is running.
If the virtual maschine xxxx is stopped, all LVM disk for this maschine are gone.

So the LVM disks on a stopped VM can not be accessed for migration to a other system.

 

[vshaulsk]

When you restart the VM on the same node, the disks return ?
- Not visible in the VM hardware configuration when VM stops ?
- Not visible under storage (Local-Storage or whatever you have it named) ?

 

[feldi]

- When you restart the VM on the same node, the disks return ? -> Yes
- Not visible in the VM hardware configuration when VM stops ? -> No, still visible
- Not visible under storage (Local-Storage or whatever you have it named) ? -> Yes

It looks like a problem with the storage mapper?

Here are some additional information:
VM 3410 is running
---------------------------------
>ls -l /dev/vg11_05
total 0
lrwxrwxrwx 1 root root 8 May 5 15:27 vm-3011-disk-1 -> ../dm-49
lrwxrwxrwx 1 root root 8 May 5 15:27 vm-3107-disk-1 -> ../dm-41
lrwxrwxrwx 1 root root 8 May 5 15:27 vm-3201-disk-1 -> ../dm-48
lrwxrwxrwx 1 root root 8 May 5 15:27 vm-3202-disk-1 -> ../dm-47
lrwxrwxrwx 1 root root 8 May 5 15:28 vm-3410-disk-1 -> ../dm-44
lrwxrwxrwx 1 root root 8 May 5 15:28 vm-3410-disk-2 -> ../dm-45

ls -l /dev/dm-4*
brw-rw---- 1 root disk 253, 41 May 5 15:28 /dev/dm-41
brw-rw---- 1 root disk 253, 44 May 10 13:04 /dev/dm-44
brw-rw---- 1 root disk 253, 45 May 10 13:04 /dev/dm-45
brw-rw---- 1 root disk 253, 46 May 10 13:04 /dev/dm-46
brw-rw---- 1 root disk 253, 47 May 10 13:04 /dev/dm-47
brw-rw---- 1 root disk 253, 48 May 10 13:01 /dev/dm-48
brw-rw---- 1 root disk 253, 49 May 10 13:04 /dev/dm-49


VM 3410 is stopping
---------------------------------
May 10 12:07:52 serv1017 pvedaemon[673792]: <root@pam> starting task UPID:serv1017:000A4A96:028103A4:5CD54D78:qmshutdown:3410:root@pam:
May 10 12:07:52 serv1017 pvedaemon[674454]: shutdown VM 3410: UPID:serv1017:000A4A96:028103A4:5CD54D78:qmshutdown:3410:root@pam:
May 10 12:07:55 serv1017 kernel: [420100.543419] vmbr0: port 17(tap3410i0) entered disabled state
May 10 12:07:56 serv1017 qmeventd[2939]: Starting cleanup for 3410
May 10 12:07:56 serv1017 qmeventd[2939]: trying to acquire lock...
May 10 12:07:56 serv1017 qmeventd[2939]: OK
May 10 12:07:56 serv1017 qmeventd[2939]: Finished cleanup for 3410
May 10 12:07:56 serv1017 pvedaemon[673792]: <root@pam> end task UPID:serv1017:000A4A96:028103A4:5CD54D78:qmshutdown:3410:root@pam: OK



VM 3410 is stopped
------------------------------------
>ls -l /dev/vg11_05
total 0
lrwxrwxrwx 1 root root 8 May 5 15:27 vm-3011-disk-1 -> ../dm-49
lrwxrwxrwx 1 root root 8 May 5 15:27 vm-3107-disk-1 -> ../dm-41
lrwxrwxrwx 1 root root 8 May 5 15:27 vm-3201-disk-1 -> ../dm-48
lrwxrwxrwx 1 root root 8 May 5 15:27 vm-3202-disk-1 -> ../dm-47

>ls -l /dev/dm-4*
brw-rw---- 1 root disk 253, 41 May 5 15:28 /dev/dm-41
brw-rw---- 1 root disk 253, 46 May 10 12:54 /dev/dm-46
brw-rw---- 1 root disk 253, 47 May 10 12:58 /dev/dm-47
brw-rw---- 1 root disk 253, 48 May 10 12:55 /dev/dm-48
brw-rw---- 1 root disk 253, 49 May 10 12:58 /dev/dm-49



In the proxmox Server View the disk are stil visible in the configuration
In the proxmox Storage View of vg05 the Disk image are still visible

>lvs | grep 3410
vm-3410-disk-1 vg11_04 -wi------- 550.00g
vm-3410-disk-1 vg11_05 -wi------- 100.00g
vm-3410-disk-2 vg11_05 -wi------- 200.00g


VM 3410 is starting again
---------------------------------------
May 10 13:04:35 serv1017 pvedaemon[673792]: <root@pam> starting task UPID:serv1017:000A60D0:028634F9:5CD55AC3:qmstart:3410:root@pam:
May 10 13:04:35 serv1017 pvedaemon[680144]: start VM 3410: UPID:serv1017:000A60D0:028634F9:5CD55AC3:qmstart:3410:root@pam:
May 10 13:04:39 serv1017 systemd[1]: Started 3410.scope.
May 10 13:04:39 serv1017 systemd-udevd[680215]: Could not generate persistent MAC address for tap3410i0: No such file or directory
May 10 13:04:40 serv1017 kernel: [423504.922768] device tap3410i0 entered promiscuous mode
May 10 13:04:40 serv1017 kernel: [423504.938971] vmbr0: port 17(tap3410i0) entered blocking state
May 10 13:04:40 serv1017 kernel: [423504.938975] vmbr0: port 17(tap3410i0) entered disabled state
May 10 13:04:40 serv1017 kernel: [423504.939193] vmbr0: port 17(tap3410i0) entered blocking state
May 10 13:04:40 serv1017 kernel: [423504.939195] vmbr0: port 17(tap3410i0) entered forwarding state
May 10 13:04:40 serv1017 pvedaemon[673792]: <root@pam> end task UPID:serv1017:000A60D0:028634F9:5CD55AC3:qmstart:3410:root@pam: OK

VM 3410 is running
---------------------------------------
ls -l /dev/vg11_05
total 0
lrwxrwxrwx 1 root root 8 May 5 15:27 vm-3011-disk-1 -> ../dm-49
lrwxrwxrwx 1 root root 8 May 5 15:27 vm-3107-disk-1 -> ../dm-41
lrwxrwxrwx 1 root root 8 May 5 15:27 vm-3201-disk-1 -> ../dm-48
lrwxrwxrwx 1 root root 8 May 5 15:27 vm-3202-disk-1 -> ../dm-47
lrwxrwxrwx 1 root root 8 May 10 13:04 vm-3410-disk-1 -> ../dm-44
lrwxrwxrwx 1 root root 8 May 10 13:04 vm-3410-disk-2 -> ../dm-45

 

[Stoiko Ivanov]

* How are you trying to migrate the machines - PVE's stack (should) make sure that the devices get activated for a migration (offline or online)?
* You can create the devicenodes by running `vgchange -ay`
* PVE disactivates the lvs in order to ensure consistency in shared LVM environments

* If you try to migrate with the PVE stack please post the command and outputs you're getting

Thanks!

 

[feldi]

How are you trying to migrate the machines - PVE's stack (should) make sure that the devices get activated for a migration (offline or online)?
-> offline, because this virtual machine is very large, 15 virtual disk, with total 12 TB

We are copy the disk with netcat
Source System:
dd if=/dev/vg11_04/vm-3002-disk-6 bs=150M | pv | nc 10.10.77.20 5555
Target System:
netcat -l -p 5555 | pv | dd of=/dev/zvol/vg01/vm-3002-disk-0 bs=150M

PVE disactivates the lvs in order to ensure consistency in shared LVM environments
-> For your point of view, the system works as designed.

For me is the problem solved!
I use now the command "vgchange -ay" and this is OK for me.

Many thanks
Thomas

 

[Stoiko Ivanov]

Glad this works for you!

Please mark the thread as '[SOLVED]' so that others know what to expect!
Thanks!