[one-users] Ubuntu shutdown/image save

[Sean Abbott]

I can't figure out why it wasn't working with your script.  I actually
modified the shutdown script to actually read the state of the VM
instead of just using a failed exit code:

<-...other code and comments>
deploy_name=$(virsh --connect $LIBVIRT_URI --readonly dominfo $deploy_id \
  | grep "Name:" | nawk '{print $2}')

count=0
deploy_state=$(virsh --connect $LIBVIRT_URI --readonly domstate
$deploy_name)
log_info "deploy state: $deploy_state"
while [ "$deploy_state" != "shut off" ]
do
    sleep $POLL_INTERVAL
    if [ "$count" -gt "$TIMEOUT" ]
    then
        error_message "Timeout reached and VM $deploy_id is still alive"
        echo "Timeout reached" >&2
        exit 1
    fi
    let count=count+$POLL_INTERVAL
    deploy_state=$(virsh --connect $LIBVIRT_URI --readonly domstate
$deploy_name)
    log_info "deploy state: $deploy_state"
done


This works, without any modifications to ubuntu 12.04.  Your timeout for
the machine shutting down is 5 minutes, the ubuntu 12.04 instance waits
60 seconds for user input, then shuts down in about 2 seconds, so it
still comes in well under the wire.

I still haven't figured out why the "wait until it doesn't return 0"
wasn't working, because when I ran it manually I was getting exit codes
of 1 almost instantly.  But.  My code above works, despite being really
ugly.  So.  {shrug}.

On 05/18/2012 04:56 AM, Jaime Melis wrote:
> Hello Sean,
>
> The shutdown script includes a timeout, when it's reached it assumes the VM is 
> still running. However it looks like the VM is indeed shutting down after that 
> timeout. I think that's why your vm is in such an ugly state.
>
> Possible fixes:
> * Disable the acpi popup in your VM [1][2]
> * Install the *server* distribution, not the desktop.
> * Try with the appliance we created [3]
> * Increase the timeout in the shutdown script [4] (not recommended!!)
>
> [1] 
> http://askubuntu.com/questions/66723/how-do-i-modify-the-power-options-in-ubuntu-11-10
> [2] 
> http://www.liberiangeek.net/2012/04/disable-shutdown-restart-logout-prompts-in-ubuntu-12-04-precise-pangolin/
> [3] http://appliances.c12g.com/Ubuntu-Server-12.04/
> [4] 
> https://github.com/OpenNebula/one/blob/one-3.4/src/vmm_mad/remotes/kvm/shutdown#L27 
> <https://github.com/OpenNebula/one/blob/one-3.4/src/vmm_mad/remotes/kvm/shutdown>
>
> Regards,
> Jaime
>
> On Thu, May 17, 2012 at 6:43 PM, Sean Abbott <seabbott at akamai.com 
> <mailto:seabbott at akamai.com>> wrote:
>
>     So, I'm attempting to create a master linux client image from which I
>     can deploy new clients easily.  I know that I'll eventually need to get
>     master contextualization.  However, before I get there, I'm having problems.
>
>     I created a template that used a raw hard disk and an ubuntu boot cd to
>     create a working ubuntu 12.04 VM.  Done, worked.
>     I used onevm saveas <vm-name> <disk-id> "ubuntu-base-image" to set up a
>     save.
>     I used onevm shutdown <vm-id> to shut it down.
>
>     At first, it appeared to work.  I was watching the vm from VNC and it
>     popped up a window asking if I wanted to shut it down (I need to
>     research how to make that window not happen in ubuntu).
>
>     However.  The VM transitioned to "unknown" state about 5 minutes later
>     (initially it was "shutdown"), and the log says that it didn't properly
>     finish shutdown:
>
>     Thu May 17 16:19:33 2012 [LCM][I]: New VM state is SHUTDOWN
>     Thu May 17 16:24:38 2012 [VMM][I]: Command execution fail:
>     /var/tmp/one/vmm/kvm/shutdown one-7 172.26.210.149 7 172.26.210.149
>     Thu May 17 16:24:38 2012 [VMM][E]: Timeout reached and VM one-7 is still alive
>     Thu May 17 16:24:38 2012 [VMM][I]: Timeout reached
>     Thu May 17 16:24:38 2012 [VMM][I]: ExitCode: 1
>     Thu May 17 16:24:38 2012 [VMM][I]: Failed to execute virtualization driver
>     operation: shutdown.
>     Thu May 17 16:24:38 2012 [VMM][E]: Error shuting down VM: Timeout reached
>     and VM one-7 is still alive
>     Thu May 17 16:24:38 2012 [LCM][I]: Fail to shutdown VM. Assuming that the VM
>     is still RUNNING (will poll VM).
>     Thu May 17 16:24:38 2012 [VMM][I]: ExitCode: 0
>     Thu May 17 16:24:38 2012 [VMM][I]: VM not found, setting to UNKNOWN
>
>
>     And, presumably because of this, my image status "locked" and there is
>     no data that has made it into the image.
>
>     Is the answer that I need to figure out how to make ubuntu do a
>     clean/fast ACPI shutdown before I can make this work, or is something
>     else afoot.  Corollary question:  has anyone managed to make ubuntu do a
>     clean ACPI shutdown with no intervention?
>
>     Thanks!
>
>     sean
>
>     _______________________________________________
>     Users mailing list
>     Users at lists.opennebula.org <mailto:Users at lists.opennebula.org>
>     http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>
>
>
>
> -- 
> Jaime Melis
> Project Engineer
> OpenNebula - The Open Source Toolkit for Cloud Computing
> www.OpenNebula.org <http://www.OpenNebula.org> | jmelis at opennebula.org 
> <mailto:jmelis at opennebula.org>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.opennebula.org/pipermail/users-opennebula.org/attachments/20120518/b8f12bf5/attachment-0003.htm>