[ovirt-users] [Centos7.1] [Ovirt 3.5.2] hosted engine, is there a way to resume deploy

[Yedidyah Bar David]

----- Original Message -----
> From: "wodel youchi" <wodel.youchi at gmail.com>
> To: "users" <Users at ovirt.org>
> Sent: Sunday, June 7, 2015 3:17:50 PM
> Subject: [ovirt-users] [Centos7.1] [Ovirt 3.5.2] hosted engine,	is there a way to resume deploy
> 
> Hi,
> 
> I tried to deploy ovirt hosted engine 3.5.2 on a Centos7.1
> 
> I messed up things with datacenter naming, I used something else than
> Default, and the result was that, after the DB welcome message between the
> engine and the hypervisor, there was in error (I didn't catch it I was using
> screen command over ssh without log :-( ),

You should be able to find it in the log, in /var/log/ovirt-hosted-engine-setup.

> and the last steps were not done,
> so I ended up
> an engine up but without the hypervisor being registred.
> 
> Is there a way to force the registration again, or will I have to deploy from
> the beginning and reinstall the VM engine?

This depends on your exact status, including whether there is anything in the
engine db and in the hosted-engine metadata file.

What's the output of 'hosted-engine --vm-status' on the host?

You can try to just run again hosted-engine --deploy, replying 'no' to 'is this an
additional host deploy?'.

If the engine db is "dirty", you can run engine-cleanup/engine-setup again inside
the engine vm, prior to trying deploy again.

There is currently no simple way to clean up the metadata file, see [1] for that.

[1] https://bugzilla.redhat.com/show_bug.cgi?id=1116469

Best,
-- 
Didi