[ovirt-users] Ovirt 4.1 testing backup and restore Self-hosted Engine

wodel youchi wodel.youchi at gmail.com
Fri Aug 25 22:56:43 UTC 2017


Hi again,

I found this article
https://keithtenzer.com/2017/05/02/rhev-4-1-lab-installation-and-configuration-guide/
I used the last section to delete the old hosted-engine storage, and it
worked, the minute I deleted the old hosted-storage the system imported the
new one and then imported the new VM-Manager into the Web admin portal.

Regards.

2017-08-25 23:15 GMT+01:00 wodel youchi <wodel.youchi at gmail.com>:

> Hi again,
>
> I redid the test again, I re-read the Self-Hosted Engine documentation,
> there is a link to a RedHat article https://access.redhat.com/
> solutions/1517683 which talks about how to remove the dead hostedEngine
> VM from the web admin portal.
>
> But the article does not talk about how to remove the old hosted engine
> storage, and this is what causes the problem.
>
> This storage is still pointing to the old iscsi disk used by the dead
> Manager, it's is down, but the new manager cannot detach it, saying that
> the storage domain doesn't exist, which is right, but how to force the
> Manager to delete it? I have no idea, I tried to remove it with REST API,
> without luck.
>
> I tried to import the new hosted storage, but the system said: the storage
> name is already in use. So I am stuck.
>
> any idea? do I have to delete it from the database? if yes how?
>
> Regards.
>
> 2017-08-25 20:07 GMT+01:00 wodel youchi <wodel.youchi at gmail.com>:
>
>> Hi,
>>
>> I was able to remove the hostedEngine VM, but I didn't succeed to remove
>> the old hostedEngine Storage domain.
>> I tried several time to remove it, but I couldn't, the VM engine goes in
>> pause mode. All I could do is to detach the hostedEngine from the
>> datacenter. I then put all the other data domains in maintenance mode, the
>> I reactivated my master data domain hoping that it will import the new
>> hostedEngine domain, but without luck.
>>
>> It seems like there is something missing in this procedure.
>>
>> Regards
>>
>> 2017-08-25 9:28 GMT+01:00 Alan Griffiths <apgriffiths79 at gmail.com>:
>>
>>> As I recall (a few weeks ago now) it was after restore, once the host
>>> had been registered in the Manager. However, I was testing on 4.0, so maybe
>>> the behaviour is slightly different in 4.1.
>>>
>>> Can you see anything in the Engine or vdsm logs as to why it won't
>>> remove the storage? Perhaps try removing the stale HostedEngine VM ?
>>>
>>> On 25 August 2017 at 09:14, wodel youchi <wodel.youchi at gmail.com> wrote:
>>>
>>>> Hi and thanks,
>>>>
>>>> But when to remove the hosted_engine storage ? During the restore
>>>> procedure or after ? Because after I couldn't do it, the manager refused to
>>>> put that storage in maintenance mode.
>>>>
>>>> Regards
>>>>
>>>> Le 25 août 2017 08:49, "Alan Griffiths" <apgriffiths79 at gmail.com> a
>>>> écrit :
>>>>
>>>>> As I recall from my testing. If you remove the old hosted_storage
>>>>> domain then the new one should get automatically imported.
>>>>>
>>>>> On 24 August 2017 at 23:03, wodel youchi <wodel.youchi at gmail.com>
>>>>> wrote:
>>>>>
>>>>>> Hi,
>>>>>>
>>>>>> I am testing the backup and restore procedure of the Self-hosted
>>>>>> Engine, and I have a problem.
>>>>>>
>>>>>> This haw I did the test.
>>>>>>
>>>>>> I have two hypervisors hosted-engine. I am used iSCSI disk for the VM
>>>>>> engine.
>>>>>>
>>>>>> I followed the procedure described in the Self-hosted Engine document
>>>>>> to execute the backup, I put the first host in maintenance mode, the I
>>>>>> create the backup and save it elsewhere.
>>>>>>
>>>>>> Then I've create a new iscsi disk, I reinstalled the first host with
>>>>>> the save IP/hostname, then I followed the restore procedure to get the
>>>>>> Manager up and running again.
>>>>>> - hosted-engine --deploy
>>>>>> - do not execute engine-setup, restore backup first
>>>>>> - execute engine-setup
>>>>>> - remove the host from the manager
>>>>>> - synchronize the restored manger with the host
>>>>>> - finalize deployment.
>>>>>>
>>>>>> all went well till this point, but I have a problem with the
>>>>>> VM-engine, it is shown as down in the admin portal. the ovirt-ha-agent
>>>>>> cannot retrieve the VM config from the shared storage.
>>>>>>
>>>>>> I think the problem, is that the hosted-engine storage domain is
>>>>>> still pointing to the old disk of the old manager and not the new one. I
>>>>>> don't know where is this information is stored, in the DB or in the
>>>>>> Manager's config files, but when I click Manager hosted-engine domain, I
>>>>>> can see the old LUN grayed and the new one (which is used by the restored
>>>>>> Manager) is not grayed.
>>>>>>
>>>>>> How can I fix this?
>>>>>>
>>>>>> Regards.
>>>>>>
>>>>>>
>>>>>> <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail> Garanti
>>>>>> sans virus. www.avast.com
>>>>>> <https://www.avast.com/sig-email?utm_medium=email&utm_source=link&utm_campaign=sig-email&utm_content=webmail>
>>>>>> <#m_-2590168438000725267_m_7057565286932519348_m_4457919776337353415_m_9177001278217562974_m_7731007781891096843_m_3883279741882476845_DAB4FAD8-2DD7-40BB-A1B8-4E2AA1F9FDF2>
>>>>>>
>>>>>> _______________________________________________
>>>>>> Users mailing list
>>>>>> Users at ovirt.org
>>>>>> http://lists.ovirt.org/mailman/listinfo/users
>>>>>>
>>>>>>
>>>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ovirt.org/pipermail/users/attachments/20170825/6489bb88/attachment.html>


More information about the Users mailing list