On Mon, Sep 28, 2015 at 1:35 PM, Bello Florent <florent.bello@ville-kourou.fr> wrote:
hi,
 
It's really not a single point of failure ? In this way i see 2 point of failure, if the hosted-engine break and if the mount server adress break. And it's not possible to migrate the engine.
 

Can you explain me why is not a point of failure for you ?

That server should be up when you start deploying hosted-engine.
When VDSM accesses that volume it gathers the whole brick list so it will then be able to use that volume also if the server you initially pointed to start the deployment process has died.
 

 

--

Florent BELLO

 

Le 28/09/2015 05:45, Simone Tiraboschi a écrit :



On Fri, Sep 25, 2015 at 5:36 PM, Bello Florent <florent.bello@ville-kourou.fr> wrote:

Hi,

I have 3 nodes with glusterfs installed. I configured replica 3 on my engine volume.

 
Hi,
on oVirt  3.6 we still not support hyper convergence that means that you cannot use the same hosts for virtualization and for storage with gluster.
 
So the correct approach is renaming your hosts to:
host 1 : "192.168.100.101 gluster1.localdomain gluster1"
host 2 : "192.168.100.102 gluster2.localdomain gluster2"
host 3 : "192.168.100.103 gluster3.localdomain gluster3"
 
On an additional host, deploy hosted-engine choosing the gluster host you prefer as your entry point: it will not be a single point of failure.
You don't really need to use CTDB to deploy oVirt hosted-engine.
 


i don't want to use ctdb for my hosted-engine, but when i start hosted-engine --deploy, use localhost:/engine or gluster.localdomain:/engine, gluster.localdomain is configured in all servers /etc/hosts like :
host 1 : "192.168.100.101 gluster.localdomain gluster"
host 2 : "192.168.100.102 gluster.localdomain gluster"
host 3 : "192.168.100.103 gluster.localdomain gluster",

the setup failed with :

--== STORAGE CONFIGURATION ==--

          During customization use CTRL-D to abort.
          Please specify the storage you would like to use (glusterfs, iscsi, fc, nfs3, nfs4)[nfs3]: glusterfs
[ INFO  ] Please note that Replica 3 support is required for the shared storage.
          Please specify the full shared storage connection path to use (example: host:/path): localhost:/engine
[WARNING] Due to several bugs in mount.glusterfs the validation of GlusterFS share cannot be reliable.
[ INFO  ] GlusterFS replica 3 Volume detected
[ ERROR ] Failed to execute stage 'Environment customization': Connection to storage server failed
[ INFO  ] Stage: Clean up
[ INFO  ] Generating answer file '/var/lib/ovirt-hosted-engine-setup/answers/answers-20150925121036.conf'
[ INFO  ] Stage: Pre-termination
[ INFO  ] Stage: Termination

I have to use obligatory ctdb or keepalived, or existing another way ?

--

Florent BELLO
Service Informatique
informatique@ville-kourou.fr
0594 22 31 22
Mairie de Kourou


_______________________________________________
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users


_______________________________________________
Users mailing list
Users@ovirt.org
http://lists.ovirt.org/mailman/listinfo/users