2

I usually install template virtual machines on a dedicated server and then copy the disk to other production servers. It has worked for years and still works today in all cases but one.

I have created a virtual domain on debian-8.1 / libvirt-1.2.9-9 / qemu-1:2.1+dfsg-12+deb8u1 / virtinst-1:1.0.1-5 and installed it with debian-9.5.

If I copy the disk to debian-8.6 / libvirt-1.2.9-9+deb8u3 / qemu-1:2.1+dfsg-12+deb8u7 / virtinst-1:1.0.1-5 and boot it, here is what I see on the vnc console:

[    0.536007] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.7.5-20140531_083030-gandalf 04/01/2014
[    0.536007] task: ffff92c7f98ff100 task.stack: ffffb69200001f80
[    0.536007] RIP: 0010:[<ffffffff9993aa54>]  [<ffffffff9993aa54>] symbol_string+0x44/0x90
[    0.536007] RSP: 0018:ffffb692802d0000  EFLAGS: 00010083
[    0.536007] RAX: 0000000000000053 RBX: ffffb692802d0001 RCX: ffff0a0Offffff05
[    0.536007] RDX: ffffffff9966066e RSI: ffffffff9966066e RDI: ffffb692802d0001
[    0.536007] RBP: ffffffff9a4b8918 R08: ffffffff99fd86fb R09: 0000000000000001
[    0.536007] R10: 000000000000000f R11: ffffffff9a4b8915 R12: ffffffff9a4b8ce0
[    0.536007] R13: ffff0a00ffffff05 R14: ffffffff99fd86fb R15: ffffffff99fd86fb
[    0.536007] FS:  00007f033ecb9700(0000) GS:ffff92c7ffc00000(0000) knlGS:000000000000000
[    0.536007] CS:  0010 DS: 0000 ES: 0000 CRO: 0000000080050033
[    0.536007] CR2: ffffb692802cfff8 CR3: 0000000039902000 CR4: 0000000000040670
[    0.536007] Stack:
[    0.536007] Call Trace:
[    0.536007] Code: 65 48 8b 04 25 28 00 00 00 48 89 84 24 10 00 00 00 31 c0 41 Of b6 00 48 8d 5c 24 01 48 89 df 3c 42 74 46 3c 66 74 Ob 3c 73 74 07 <e8> b7 97 dc ff eb 05 e8 40 98 dc ff 4c 89 e9 48 89 da 4c 89 e6
[    0.536007] Kernel panic - not syncing: Machine halted.
[    0.536007] Kernel Offset: 0x18600000 from 0xffffffff81000000 (relocation range: 0xffffffff80000000-0xffffffffbfffffff)
[    0.536007] ---[ end Kernel panic - not syncing: Machine halted.

I have no idea what this means and can't find any relevant Google information about it.

Does anyone knows what could be wrong?

--- EDIT -------

I have tested the same disk on various virtualization hosts. It works everywhere but on one system:

HOSTNAME  / STAT  / DEBIAN     / KERNEL         / LIBVIRT                / QEMU                      / VIRTINST
pandora     OK      debian-8.1 / 3.16.0-4-amd64 / libvirt-1.2.9-9        / qemu-1:2.1+dfsg-12+deb8u1 / virtinst-1:1.0.1-5
20200107    OK      debian-8.2 / 3.16.0-4-amd64 / libvirt-1.2.9-9+deb8u5 / qemu-1:2.1+dfsg-12+deb8u7 / virtinst-1:1.0.1-5
20200138    OK      debian-8.2 / 3.16.0-4-amd64 / libvirt-1.2.9-9+deb8u5 / qemu-1:2.1+dfsg-12+deb8u7 / virtinst-1:1.0.1-5
20200121    OK      debian-8.5 / 3.16.0-4-amd64 / libvirt-1.2.9-9+deb8u3 / qemu-1:2.1+dfsg-12+deb8u6 / virtinst-1:1.0.1-5
20200176    ERR     debian-8.6 / 3.16.0-4-amd64 / libvirt-1.2.9-9+deb8u3 / qemu-1:2.1+dfsg-12+deb8u7 / virtinst-1:1.0.1-5
20200046    OK      debian-9.5 / 4.9.0-7-amd64  / libvirt-3.0.0-4+deb9u3 / qemu-1:2.8+dfsg-6+deb9u4  / virtinst-1:1.4.0-5
20200084    OK      debian-9.5 / 4.9.0-7-amd64  / libvirt-3.0.0-4+deb9u3 / qemu-1:2.8+dfsg-6+deb9u4  / virtinst-1:1.4.0-5
20200114    OK      debian-9.5 / 4.9.0-7-amd64  / libvirt-3.0.0-4+deb9u3 / qemu-1:2.8+dfsg-6+deb9u4  / virtinst-1:1.4.0-5

Thanks for your help
Santiago

Chebarbudo
  • 73
  • 1
  • 6
  • Where is the rest of it? – Michael Hampton Nov 09 '18 at 17:07
  • The rest of what? – Chebarbudo Nov 09 '18 at 17:23
  • The rest of the kernel panic message, of course. Some of it is missing. – Michael Hampton Nov 09 '18 at 17:27
  • Ah well. The boot goes so fast that this is the only part I can read. I boot the vm from host, connect to vnc console and too late, this is all I see. Is there a way to see more ? to scroll up ? – Chebarbudo Nov 09 '18 at 17:38
  • You could use a serial console, or make the virtual framebuffer larger. but based on your tests it's probably a KVM bug that was present in a specific kernel version, has since been fixed, and the solution is to upgrade your KVM hosts. – Michael Hampton Nov 09 '18 at 17:56
  • Michael, I want to thank you for the time taken to help me. Although I don't believe in the kernel version problem. I have now tested the disk on 8 different hosts (see my edit) and still only one fails to launch. It has the same kernel version as others, and also the same libvirt, qemu and virtinst version. I don't see what can be wrong. I'm trying to find out how to use a serial console with qemu (if you have a hint, I'm interested). – Chebarbudo Nov 09 '18 at 20:19
  • Have you considered it might be a physical hardware problem, then? – Michael Hampton Nov 09 '18 at 20:23
  • And of course you should update those oldstable hosts with [LTS](https://wiki.debian.org/LTS/Using) or upgrade them to stable. – Michael Hampton Nov 09 '18 at 20:37
  • Physical hardware problem is unlikely because I have other VMs running smoothly on the same buggy host. – Chebarbudo Nov 09 '18 at 20:50

0 Answers0