[ovirt-users] high availability

[Nir Soffer]

On Fri, Feb 3, 2017 at 9:49 PM, Yaniv Kaul <ykaul at redhat.com> wrote:
>
>
> On Fri, Feb 3, 2017 at 8:42 PM, cmc <iucounu at gmail.com> wrote:
>>
>> Hi,
>>
>> I have some questions about oVirt's high availability features for
>> VMs. My understanding is that it relies on the engine host to monitor
>> and manage the hypervisor hosts, so that in the case of a
>> unrecoverable failure of one those hosts, it will fence the host and
>> migrate any VM that is designated as highly available to another host
>> in the cluster. However, if the engine is itself hosted as a VM on a
>> host that fails, this process cannot take place, as the engine will be
>> down and cannot initiate monitoring, fencing and migration - is that
>> correct?
>
>
> The hosted-engine has its own HA mechanism.

You may find this useful:
http://www.ovirt.org/documentation/self-hosted/Self-Hosted_Engine_Guide/

> In addition, in 4.1 we are introducing a feature which allows HA without
> fencing, in a similar manner to hosted-engine - by a lock on the storage
> side.

For more info on this see:
https://www.ovirt.org/develop/release-management/features/storage/vm-leases/

Nir

> Y.
>
>>
>> There is the option of hosting the engine externally on dedicated
>> hardware, or on another cluster, but then it is still a single point
>> of failure. I recall reading about plans for an HA engine in the
>> future though.
>>
>> Can someone tell me what the roadmap is? Is there a plan to put
>> something like an HA agent on all the hypervisors in the cluster so
>> there is no single point of failure?
>>
>> Thanks for any information,
>>
>> Cam
>> _______________________________________________
>> Users mailing list
>> Users at ovirt.org
>> http://lists.ovirt.org/mailman/listinfo/users
>
>
>
> _______________________________________________
> Users mailing list
> Users at ovirt.org
> http://lists.ovirt.org/mailman/listinfo/users
>