[ovirt-users] Testing self hosted engine in 3.6: hostname not resolved error

Yedidyah Bar David didi at redhat.com
Wed Oct 21 14:08:59 UTC 2015


On Wed, Oct 21, 2015 at 5:00 PM, Gianluca Cecchi
<gianluca.cecchi at gmail.com> wrote:
> Hello,
> playing with an environment where I have a CentOS 7.1 + updates server with
> 3.6 repos configured.
> no DNS resolution in place.
> running the command
> hosted-engine --deploy
>
> I initially get, during the prompts inside setup, the message if I want to
> set hostname for the sh engine and if I want to setup /etc/hosts of
> hypervisor

IIRC we only suggest to update /etc/hosts of the engine, not the host.

> and sh engine so that it will add them with clud-init, that
> translates in these lines inside log file:
>
>
> 2015-10-21 15:45:29 DEBUG otopi.context context.dumpEnvironment:510 ENV
> OVEHOSTED_VM/cloudinitInstanceDomainName=str:'localdomain.local'
> 2015-10-21 15:45:29 DEBUG otopi.context context.dumpEnvironment:510 ENV
> OVEHOSTED_VM/cloudinitInstanceHostName=str:'shengine.localdomain.local'
> 2015-10-21 15:45:29 DEBUG otopi.context context.dumpEnvironment:510 ENV
> OVEHOSTED_VM/cloudinitRootPwd=str:'**FILTERED**'
> 2015-10-21 15:45:29 DEBUG otopi.context context.dumpEnvironment:510 ENV
> OVEHOSTED_VM/cloudinitVMDNS=str:'192.168.122.1'
> 2015-10-21 15:45:29 DEBUG otopi.context context.dumpEnvironment:510 ENV
> OVEHOSTED_VM/cloudinitVMETCHOSTS=bool:'True'
>
> But then I get the error related to failure of sh engine hostname resolution
> and installation terminates, with these finale lines in log file.....
>
> 2015-10-21 15:45:29 DEBUG otopi.context context.dumpEnvironment:510 ENV
> SYSTEM/reboot=bool:'False'
> 2015-10-21 15:45:29 DEBUG otopi.context context.dumpEnvironment:510 ENV
> SYSTEM/rebootAllow=bool:'True'
> 2015-10-21 15:45:29 DEBUG otopi.context context.dumpEnvironment:510 ENV
> SYSTEM/rebootDeferTime=int:'10'
> 2015-10-21 15:45:29 DEBUG otopi.context context.dumpEnvironment:514
> ENVIRONMENT DUMP - END
> 2015-10-21 15:45:29 DEBUG otopi.context context._executeMethod:142 Stage
> pre-terminate METHOD otopi.plugins.otopi.dialog.cli.Plugin._pre_terminate
> 2015-10-21 15:45:29 DEBUG otopi.context context._executeMethod:148 condition
> False
> 2015-10-21 15:45:29 INFO otopi.context context.runSequence:427 Stage:
> Termination
> 2015-10-21 15:45:29 DEBUG otopi.context context.runSequence:431 STAGE
> terminate
> 2015-10-21 15:45:29 DEBUG otopi.context context._executeMethod:142 Stage
> terminate METHOD otopi.plugins.otopi.dialog.human.Plugin._terminate
> 2015-10-21 15:45:29 DEBUG otopi.context context._executeMethod:142 Stage
> terminate METHOD otopi.plugins.otopi.dialog.machine.Plugin._terminate
> 2015-10-21 15:45:29 DEBUG otopi.context context._executeMethod:148 condition
> False
> 2015-10-21 15:45:29 DEBUG otopi.context context._executeMethod:142 Stage
> terminate METHOD otopi.plugins.otopi.core.log.Plugin._terminate
>
> and the message of a generated answer file.
>
> Why does it ask and then fails?

Generally speaking, you are supposed to take care of all name resolution
yourself. Changing /etc/hosts on the engine vm was added mainly to allow
unattended setup of it.

Note that you also have to make sure the engine can resolve the host, and
that the hosts can resolve each other. See also [1] about that.

[1] https://bugzilla.redhat.com/show_bug.cgi?id=1188675

> Can I re-run using the answer file? In this case do I have to pre-insert sh
> engine hostname inside /etc/hosts of hypervisor?

Yes.

Best regards,

>
> Thanks.
> Gianluca
>
> _______________________________________________
> Users mailing list
> Users at ovirt.org
> http://lists.ovirt.org/mailman/listinfo/users
>



-- 
Didi



More information about the Users mailing list