← Back to team overview

openstack team mailing list archive

Re: [OPENSTACK] VMs don't start on host machine reboot (Grizzly over Ubuntu 13.04)

 

An update:
I used the glance image-create command to add the image to the glance DB.
If I use this image for creating VM, then the problem of VM not booting on
host machine node is gone away. I am still puzzled as to what is the
difference there it is vs. the original instance xml file. are there any
other files that I should be comparing with ?

Output of command "qemu-img info
/var/lib/nova/instances/6d6b2f12-0f59-4b94-adc2-b527d8bf4d94/disk" ::
+++++++++++++++++++++++++++++++++++++++++++++++
image: /var/lib/nova/instances/6d6b2f12-0f59-4b94-adc2-b527d8bf4d94/disk
file format: qcow2
virtual size: 160G (171798691840 bytes)
disk size: 200K
cluster_size: 65536
backing file:
/var/lib/nova/instances/_base/f0e93d9c7faf165914420bff74264d92be2bd84a
root@osk-team:/etc/libvirt/qemu# qemu-img info
/var/lib/nova/instances/6d6b2f12-0f59-4b94-adc2-b527d8bf4d94/disk.local
image:
/var/lib/nova/instances/6d6b2f12-0f59-4b94-adc2-b527d8bf4d94/disk.local
file format: qcow2
virtual size: 160G (171798691840 bytes)
disk size: 32G
cluster_size: 65536
backing file: /var/lib/nova/instances/_base/ephemeral_160_default
+++++++++++++++++++++++++++++++++++++++++++++++


The instance specific section in file
/etc/libvirt/qemu/instance-0000001c.xml is :
============================================================

  <devices>
    <emulator>/usr/bin/kvm</emulator>
    <disk type='file' device='cdrom'>
      <driver name='qemu' type='qcow2' cache='none'/>
      <source
file='/var/lib/nova/instances/6d6b2f12-0f59-4b94-adc2-b527d8bf4d94/disk'/>
      <target dev='hda' bus='ide'/>
      <readonly/>
      <address type='drive' controller='0' bus='0' target='0' unit='0'/>
    </disk>
    <disk type='file' device='disk'>
      <driver name='qemu' type='qcow2' cache='none'/>
      <source
file='/var/lib/nova/instances/6d6b2f12-0f59-4b94-adc2-b527d8bf4d94/disk.local'/>
      <target dev='vda' bus='virtio'/>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x04'
function='0x0'/>
    </disk>
    <controller type='usb' index='0'>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x01'
function='0x2'/>
    </controller>
    <controller type='ide' index='0'>
      <address type='pci' domain='0x0000' bus='0x00' slot='0x01'
function='0x1'/>
    </controller>
==============================================================
-Nikhil


On Wed, May 15, 2013 at 3:34 PM, Nikhil Mittal <mittalnikhil@xxxxxxxxx>wrote:

> Some additional info: I tried the host machine reboot after installing a
> qcow2 format ubuntu image as VM. And it works fine i.e. after power-cycling
> the host node, the VM is able to boot again. I suspect either there is bug
> with Grizzly release in handling certain type of ISO files or the ISO needs
> to be built keeping certain things in mind. FYI, the ISO image that I am
> using is an in-house custom built image which we have been using
> successfully across multiple platforms as well as on Openstack Essex
> release.
>
>
> On Wed, May 15, 2013 at 2:02 PM, Nikhil Mittal <mittalnikhil@xxxxxxxxx>wrote:
>
>> Output of command "ls -l
>> /var/lib/nova/instances/_base/6a2ba9ce1da1fbdc28a3da86332cb65193a2bf96"
>>
>> -rw-r--r-- 1 nova kvm 6285240320 May 15 06:21
>> /var/lib/nova/instances/_base/6a2ba9ce1da1fbdc28a3da86332cb65193a2bf96
>>
>> Seems normal to me...
>>
>> I added the image in glance using command:
>> glance add name="CCM_9001" is_public=true container_format=bare
>> disk_format=iso < *image_file*
>>
>> The same commands have been working fine in Essex release.
>>
>>
>> On Wed, May 15, 2013 at 12:51 PM, lzy.dev@xxxxxxxxx <lzy.dev@xxxxxxxxx>wrote:
>>
>>> Can you check the backing file:
>>> /var/lib/nova/instances/_base/6a2ba9ce1da1fbdc28a3da86332cb65193a2bf96
>>> is in normal?
>>>
>>> On Wed, May 15, 2013 at 3:16 PM, Nikhil Mittal <mittalnikhil@xxxxxxxxx>
>>> wrote:
>>> > Any suggestions on debugging this issue ? As I mentioned below, the
>>> command
>>> > qemu-img shows information where the file format is qcow2 whereas the
>>> image
>>> > used in this instance is ISO. Is this a bug or am I missing something ?
>>> > Thanks,
>>> > -Nikhil
>>> >
>>> >
>>> > On Tue, May 14, 2013 at 3:30 PM, Nikhil Mittal <mittalnikhil@xxxxxxxxx
>>> >
>>> > wrote:
>>> >>
>>> >> Here it is:
>>> >> ===============
>>> >> image:
>>> /var/lib/nova/instances/a562a6c9-9253-4f22-8076-344e855f713d/disk
>>> >> file format: qcow2
>>> >> virtual size: 160G (171798691840 bytes)
>>> >> disk size: 200K
>>> >> cluster_size: 65536
>>> >> backing file:
>>> >> /var/lib/nova/instances/_base/6a2ba9ce1da1fbdc28a3da86332cb65193a2bf96
>>> >> ===============
>>> >> BTW, I launched the instance with ISO image format NOT qcow2. Don't
>>> know
>>> >> why it is showing the format as qcow2.
>>> >>
>>> >>
>>> >> On Tue, May 14, 2013 at 3:23 PM, lzy.dev@xxxxxxxxx <lzy.dev@xxxxxxxxx
>>> >
>>> >> wrote:
>>> >>>
>>> >>> Nikhil
>>> >>>
>>> >>> What  the 'qemu-img info
>>> >>> /var/lib/nova/instances/a562a6c9-9253-4f22-8076-344e855f713d/disk'
>>> >>> command say?
>>> >>>
>>> >>> Zhi Yan
>>> >>>
>>> >>> On Tue, May 14, 2013 at 5:43 PM, Nikhil Mittal <
>>> mittalnikhil@xxxxxxxxx>
>>> >>> wrote:
>>> >>> > Wang,
>>> >>> > I guess you are referring to to the instance specific XML file
>>> under
>>> >>> > /etc/libvirt/qemu. It has the following section having 'vda':
>>> >>> >
>>> >>> > ===========================
>>> >>> > <devices>
>>> >>> >     <emulator>/usr/bin/kvm</emulator>
>>> >>> >     <disk type='file' device='disk'>
>>> >>> >       <driver name='qemu' type='qcow2' cache='none'/>
>>> >>> >       <source
>>> >>> >
>>> >>> >
>>> file='/var/lib/nova/instances/a562a6c9-9253-4f22-8076-344e855f713d/disk'/>
>>> >>> >       <target dev='vda' bus='virtio'/>
>>> >>> >       <address type='pci' domain='0x0000' bus='0x00' slot='0x04'
>>> >>> > function='0x0'/>
>>> >>> >     </disk>
>>> >>> > ....
>>> >>> > ===========================
>>> >>> > Can you specify what exactly I need to change ?
>>> >>> > Thanks,
>>> >>> > Nikhil
>>> >>> >
>>> >>> >
>>> >>> >
>>> >>> > On Tue, May 14, 2013 at 1:45 PM, Wangpan <
>>> hzwangpan@xxxxxxxxxxxxxxxx>
>>> >>> > wrote:
>>> >>> >>
>>> >>> >> seems like the roog disk is not a bootable device, you may need to
>>> >>> >> check
>>> >>> >> the vda/xvda device in libvirt XML config file is bootable or not.
>>> >>> >>
>>> >>> >> 2013-05-14
>>> >>> >> ________________________________
>>> >>> >> Wangpan
>>> >>> >> ________________________________
>>> >>> >> 发件人:Nikhil Mittal
>>> >>> >> 发送时间:2013-05-14 15:59
>>> >>> >> 主题:[Openstack] [OPENSTACK] VMs don't start on host machine reboot
>>> >>> >> (Grizzly
>>> >>> >> over Ubuntu 13.04)
>>> >>> >> 收件人:"openstack"<openstack@xxxxxxxxxxxxxxxxxxx>
>>> >>> >> 抄送:
>>> >>> >>
>>> >>> >> Hello,
>>> >>> >> I recently installed Grizzly (single-node) on Ubuntu 13.04 on x86
>>> 64
>>> >>> >> bit
>>> >>> >> having two physical NICs. I posted about the networking issues
>>> that I
>>> >>> >> was
>>> >>> >> facing after the installation in an earlier post.
>>> >>> >> Well, while I was debugging the networking issue, I came across a
>>> more
>>> >>> >> serious issue -- my host machine got rebooted (power-cycle) and on
>>> >>> >> reboot I
>>> >>> >> see that my two VMs are shown in "shut-off" Status and "shut-down"
>>> >>> >> Power
>>> >>> >> state in the Dashboard GUI. The options available to me are only
>>> "soft
>>> >>> >> reboot", "hard reboot", "terminate instance", "Disassociate
>>> Floating
>>> >>> >> IP" and
>>> >>> >> "Edit Instance". I did first "soft reboot" and the console shows
>>> "No
>>> >>> >> bootable device". Then i tried "Hard reboot" and again the same
>>> >>> >> message "No
>>> >>> >> bootable device".
>>> >>> >> I checked that all services are running fine and there are no
>>> error
>>> >>> >> messages in the logs for nova, quantum, etc.
>>> >>> >>
>>> >>> >> I followed installation instructions based on the following link:
>>> >>> >>
>>> >>> >>
>>> >>> >>
>>> https://github.com/mseknibilel/OpenStack-Grizzly-Install-Guide/blob/OVS_SingleNode/OpenStack_Grizzly_Install_Guide.rst
>>> >>> >>
>>> >>> >> Thanks,
>>> >>> >> Nikhil
>>> >>> >>
>>> >>> >
>>> >>> >
>>> >>> > _______________________________________________
>>> >>> > Mailing list: https://launchpad.net/~openstack
>>> >>> > Post to     : openstack@xxxxxxxxxxxxxxxxxxx
>>> >>> > Unsubscribe : https://launchpad.net/~openstack
>>> >>> > More help   : https://help.launchpad.net/ListHelp
>>> >>> >
>>> >>
>>> >>
>>> >
>>>
>>
>>
>

References