My setup 1 engine Fedora 19 ovirt 3.4 managing 3 nodes Fedora 19 + ovirt 3.4 updated; those nodes acts also as gluster server for 2 storage domains, each domain 3 bricks replicated  ( so 3 bricks each).

I have one questions regarding the fields one have to complete when defining a new storage domain and its importance.

First, field "Use Host" - you have to complete with one of the nodes. Of course you can use any of the host in cluster, but which is its significance/importance?  If you reboot that node, you suppose to have any issue with the storage?

The same question with field "Path" - gluster case - "Path to gluster volume to mount" formatted as "HOST:VOLUME" , where host can be the same as above mentioned field-  "Use Host" or any af the gluster cluster member; which is its significance/importance? Again If you reboot that node, you suppose to have any issue with the storage?

My experience with rebooting these nodes, after moving VM's and putting it in Maintenance was good, until yesterday, when after such reboot I've got some corrupted VM images, and subsequently  I had to restore them.

Error message follows :

Exit message: cannot read header '/rhev/data-center/mnt/glusterSD/10.125.1.194:gluster__data3/5fe658b1-43ff-4bfd-97d6-28217cf415df/images/321ad473-3741-4825-8ac0-6c416aa8f490/1a21454d-9931-4ed8-ba38-7126528664a1': Input/output error."

Did I miss some step(s), beside moving VM's and puting node in maintenace, before reboot one host node, servig also as gluster node?

Any hint, link would be appreciated!

Thanks!