[Bug 1013782] [NEW] Stop/start a KVM instance with volumes attached produces an error state

[Ryan Finnie]

Public bug reported:

When a running instance with an attached volume is stopped and then
started, the instance refuses to boot and goes into an error state.
This appears to be caused by nova-compute incorrectly building the
libvirt.xml file.

2012-06-15 03:54:14 TRACE nova.compute.manager [instance: 972b355b-21cc-
4ca8-bbb7-67af1bf2ee7f] libvirtError: internal error Invalid harddisk
device name: /dev/vdz

A log of the start and the generated libvirt.xml are attached.  In
particular:

            <disk type='block'>
                     <driver name='qemu' type='raw' cache='none'/>
                     <source dev='/dev/disk/by-path/ip-10.55.61.15:3260-iscsi-iqn.2010-10.org.openstack:volume-00000053-lun-1'/>
                     <target dev='/dev/vdz' bus='virtio'/>
                 </disk>

libvirt expects <target dev='vdz' bus='virtio'/>, and refuses to boot
otherwise.  Issuing a reboot to the instance does not trigger this (and
the XML file is not even updated with the volume, presumably because the
process is simply given a hard reboot order).

Version: 2012.1-0ubuntu2.2 (with SRU patch from 2012.1-0ubuntu2.3
manually applied)

** Affects: nova (Ubuntu)
     Importance: Undecided
         Status: New


** Tags: canonistack

-- 
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to nova in Ubuntu.
https://bugs.launchpad.net/bugs/1013782

Title:
  Stop/start a KVM instance with volumes attached produces an error
  state

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/nova/+bug/1013782/+subscriptions