[one-users] ebtables not deactived at "onevm delete"

[Shi Jin]

Thanks.
But the same problem happens to "onevm shutdown". Here the vm.log
Mon Oct  5 16:27:50 2009 [LCM][I]: New VM state is SHUTDOWN
Mon Oct  5 16:28:26 2009 [LCM][I]: New VM state is EPILOG
Mon Oct  5 16:28:31 2009 [HKM][I]: Command execution fail:
'/srv/cloud/one/share/hooks/ebtables-kvm stop one-19'
Mon Oct  5 16:28:31 2009 [HKM][I]: STDERR follows.
Mon Oct  5 16:28:31 2009 [HKM][I]: Connecting to uri: qemu:///system
Mon Oct  5 16:28:31 2009 [HKM][I]: error: failed to get domain 'one-19'
Mon Oct  5 16:28:31 2009 [HKM][I]: error: Domain not found: no domain
with matching name 'one-19'
Mon Oct  5 16:28:31 2009 [HKM][I]:
/srv/cloud/one/share/hooks/ebtables-kvm:21: undefined method
`elements' for nil:NilClass (NoMethodError)
Mon Oct  5 16:28:31 2009 [HKM][I]: ExitCode: 1
Mon Oct  5 16:28:31 2009 [HKM][E]: Error executing Hook: ebtables-shutdown.
Mon Oct  5 16:28:31 2009 [TM][I]: tm_delete.sh: Deleting /opt/cloud/VM/19/images
Mon Oct  5 16:28:31 2009 [TM][I]: tm_delete.sh: Executed "ssh node1 rm
-rf /opt/cloud/VM/19/images".
Mon Oct  5 16:28:31 2009 [DiM][I]: New VM state is DONE

Again the reason seems to be that the script executed after the VM is destroyed.
Could there be a choice of running a script before the shutdown action?
Thanks.

Shi

On Mon, Oct 5, 2009 at 4:36 AM, Javier Fontan <jfontan at gmail.com> wrote:
> Hello,
>
> As delete is meant to be used only in cases where the VM is not
> behaving or had troubles, not a common case, we had not added hooks to
> this state change. Doing so requires several changes in the core and
> we do not want to modify it until 1.4.0. We will look into that after
> the release is made.
>
> Bye
>
> On Sun, Oct 4, 2009 at 4:41 AM, Shi Jin <jinzishuai at gmail.com> wrote:
>> Hi there,
>> I just found that when "onevm delete" is excuted, the ebtables hook was not
>> run properly.
>> I think this maybe due to that fact that the hook is run after the virtual
>> machine is destroy. However, the hook script needs the to run virsh to get
>> the VM property, which at that point is no longer alive. Can this be fixed?
>> Thanks.
>> --
>> Shi Jin, Ph.D.
>>
>>
>> _______________________________________________
>> Users mailing list
>> Users at lists.opennebula.org
>> http://lists.opennebula.org/listinfo.cgi/users-opennebula.org
>>
>>
>
>
>
> --
> Javier Fontan, Grid & Virtualization Technology Engineer/Researcher
> DSA Research Group: http://dsa-research.org
> Globus GridWay Metascheduler: http://www.GridWay.org
> OpenNebula Virtual Infrastructure Engine: http://www.OpenNebula.org
>



-- 
Shi Jin, Ph.D.