Hi,
I have experimented on the issue and figured out the reason for the
original issue.
You are right, that the vm1 is not properly stopped. This is due to the
known issue in the graceful shutdown introduced in the ovirt 4.2.
The vm on the host in shutdown are killed, but are not marked as stopped.
This results in the behavior you have observed.
Luckily, the patch is already done and present in the latest ovirt.
However, be ware that gracefully shutting down the host, will result in
graceful shutdown of
the VMs. This result in engine not migrating them, since they have been
terminated gracefully.
Hope this helps.
Best,
Petr
On Fri, Feb 2, 2018 at 6:00 PM, Simone Tiraboschi <stirabos(a)redhat.com>
wrote:
On Thu, Feb 1, 2018 at 1:06 PM, Pym <pym0914(a)163.com> wrote:
> The environment on my side may be different from the link. My VM1 can be
> used normally after it is started on host2, but there is still information
> left on host1 that is not cleaned up.
>
> Only the interface and background can still get the information of vm1 on
> host1, but the vm2 has been successfully started on host2, with the HA
> function.
>
> I would like to ask a question, whether the UUID of the virtual machine
> is stored in the database or where is it maintained? Is it not successfully
> deleted after using the HA function?
>
>
I just encounter a similar behavior:
after a reboot of the host 'vdsm-client Host getVMFullList' is still
reporting an old VM that is not visible with 'virsh -r list --all'.
I filed a bug to track it:
https://bugzilla.redhat.com/show_bug.cgi?id=1541479
>
>
>
>
> 2018-02-01 16:12:16,"Simone Tiraboschi" <stirabos(a)redhat.com> :
>
>
>
> On Thu, Feb 1, 2018 at 2:21 AM, Pym <pym0914(a)163.com> wrote:
>
>>
>> I checked the vm1, he is keep up state, and can be used, but on host1
>> has after shutdown is a suspended vm1, this cannot be used, this is the
>> problem now.
>>
>> In host1, you can get the information of vm1 using the "vdsm-client Host
>> getVMList", but you can't get the vm1 information using the "virsh
list".
>>
>>
> Maybe a side effect of
https://bugzilla.redhat.com
> /show_bug.cgi?id=1505399
>
> Arik?
>
>
>
>>
>>
>>
>> 2018-02-01 07:16:37,"Simone Tiraboschi" <stirabos(a)redhat.com> :
>>
>>
>>
>> On Wed, Jan 31, 2018 at 12:46 PM, Pym <pym0914(a)163.com> wrote:
>>
>>> Hi:
>>>
>>> The current environment is as follows:
>>>
>>> Ovirt-engine version 4.2.0 is the source code compilation and
>>> installation. Add two hosts, host1 and host2, respectively. At host1, a
>>> virtual machine is created on vm1, and a vm2 is created on host2 and HA is
>>> configured.
>>>
>>> Operation steps:
>>>
>>> Use the shutdown -r command on host1. Vm1 successfully migrated to
>>> host2.
>>> When host1 is restarted, the following situation occurs:
>>>
>>> The state of the vm2 will be shown in two images, switching from up and
>>> pause.
>>>
>>> When I perform the "vdsm-client Host getVMList" in host1, I will
get
>>> the information of vm1. When I execute the "vdsm-client Host
getVMList" in
>>> host2, I will get the information of vm1 and vm2.
>>> When I do "virsh list" in host1, there is no virtual machine
>>> information. When I execute "virsh list" at host2, I will get
information
>>> of vm1 and vm2.
>>>
>>> How to solve this problem?
>>>
>>> Is it the case that vm1 did not remove the information on host1 during
>>> the migration, or any other reason?
>>>
>>
>> Did you also check if your vms always remained up?
>> In 4.2 we have libvirt-guests service on the hosts which tries to
>> properly shutdown the running VMs on host shutdown.
>>
>>
>>>
>>> Thank you.
>>>
>>>
>>>
>>>
>>> _______________________________________________
>>> Users mailing list
>>> Users(a)ovirt.org
>>>
http://lists.ovirt.org/mailman/listinfo/users
>>>
>>>
>>
>>
>>
>>
>
>
>
>
>
_______________________________________________
Users mailing list
Users(a)ovirt.org
http://lists.ovirt.org/mailman/listinfo/users